14053636495342d22ce2ba3a93feba38f8cf4aab
88 Commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
 vh
|
b8435ceb6f |
Set mog-sec's context to a measured ceiling so it refuses instead of dying
Operator: "loading up the context killed sec again." That reproducer is what finally made the failure legible, and it showed the previous four fixes had all been aimed at the wrong quantity. What the KV pool can hold and what the card can process at depth are different numbers. Cutting context 420k to 384k to 320k, pinning the KV in bytes, and dropping the prefill chunk from 16384 to 4096 all sized the pool. The crashes were governed by the transient needed to process a prefill chunk against a quarter million tokens of resident KV, which scales with depth and not with pool size. Each change helped and none fixed it. Bisected against the real reproducer, with a non-repeating prompt because prefix caching would let a repeated one hash to cached blocks and never prefill deep: 113,247 prompt tokens SURVIVED (27 s) 200,088 prompt tokens SURVIVED (174 s) ~285,000 prompt tokens ENGINE DIED, HTTP 500, container restarted The sustainable ceiling therefore sits between 200k and 285k with gen idle, and gen shares the card with its load uncontrolled, so 163,840 takes about 20% margin under the proven-good depth rather than sitting at the measured edge. The ceiling's purpose is the refusal. Verified after the change: a 149,073-token request serves in 41 s, and requests at both 200k and the ~285k depth that killed the engine now return a clean 400 naming the limit in under a second with the seat untouched. A seat that refuses what it cannot serve is strictly better than one that dies trying. Concurrency went 1.03x to 2.09x. The compose header's "served at native 262K" was never actually deliverable on a shared card; it had simply not been exercised at depth until today. The probe is committed rather than described, so the ceiling can be re-measured when the card's tenancy changes. |
||
|
vh
|
9a916a759f |
Swap the MeroMero A4B onto the erp-seat seat as char-rp-fast, retire the Pfish-6 alias
Operator: "replace that a4b moe over pfish-6 -- remove the pfish-6 alias and create an alias for char-rp-fast." G4-MeroMero-26B-A4B-it-uncensored-heretic-NVFP4A16 is live on ana-ml2 :8021 under its own served name, behind the new gateway alias char-rp-fast. Pfish-6 is gone from the gateway and now returns an explicit 400 rather than a substitution; 0 of 17 LiteLLM keys scoped it, so nothing was orphaned. The compose project name stays erp-seat because asset-engine derives seat liveness from it. The first quant of that A4B served NaN and passed its healthcheck doing it. It was built with the dense v2-31B recipe, whose ignore list has no router regex, so all 30 MoE routers were quantized to 4 bits -- and a 4-bit router changes which experts run rather than degrading them. Quant rc=0, healthcheck green, correct KV pool, correct served name, and every completion returned finish_reason=length with the full token count and content: null. The model was emitting a full budget of tokens that decoded to the empty string. Raw /v1/completions was empty too, ruling out the chat template and the reasoning parser. The signal that named it was logprobs: vLLM refused to serialize the response, "Out of range float values are not JSON compliant: nan". The lesson is about the control rather than the router. That tree had already been structurally diffed and passed -- against a verified-good DENSE quant of the same Gemma-4 family. A dense model has no routers, so the single thing that was wrong was the single thing the control could not distinguish. Diffing instead against Pfish-6, a known-good quant of the same 26B-A4B MoE, gave it in one line: 222 ignore entries against 252, the 30 missing being layers.N.router.proj. A positive control is only worth what it can distinguish, and "same family" is not "same architecture class". Re-quantized with the MoE recipe, whose dry-run asserts 11,520 expert Linears and refuses a router in the quantize set before any GPU time. The live seat then passed prose with no channel-prefix leak, a solid-colour image read correctly, an auto tool call parsed, finite logprobs, and KV 534,649 tokens / 2.04x carried over from Pfish-6 unchanged. The broken tree is parked on ana-ml2 as ...-NVFP4A16.BROKEN-routers-quantized-20260910. Section 4.4's temp port was not reachable: 15.9 GiB of weights plus KV plus multimodal encoder-cache profiling does not fit in the ~19 GiB free beside GPU1's six other tenants -- 0.20 utilization refused admission, 0.185 OOM'd in encoder profiling. The substitute was reversibility and ordering: named .env backup, prove the seat on its real port while no alias points at it, move the alias last. That is why a NaN-serving seat never reached a consumer. The seat was down about 16 minutes across two attempts; no consumer saw a broken alias. Playbook gains the router-quant failure signature and the control-class rule in 3.15, and a logprobs check in 4.4. seat_verify.py carries that check as check 6. Quality is NOT established: no RP eval, no long-context check, no A/B against Pfish-6 or char-rp. Samplers are the author's card values, untuned here. |
||
|
vh
|
1a5bc2ddf1 |
Land the MeroMero v2-31B NVFP4A16 quant and record the pinned-transformers trap
The v2 dense quant had failed four times. Attempt 5 lands it at 19 G. The blocker was not what it looked like. `AmbiguousGlobalPerLayerAttributeError` on `head_dim` read as a malformed upload -- DogOnKeyboard's config carries a `per_layer_config` key zerofata's canonical one lacks -- and the standing fix was to force `allow_global_per_layer_attribute_access=True`. Both halves were wrong. `pip install llmcompressor==0.13.0` downgrades transformers 5.16.1 -> 5.14.1. The config was serialized by 5.16.1, which materializes `per_layer_config` from `global_head_dim` + `layer_types`; 5.14.1 carries the heterogeneity guard but not the gemma4 resolver. Under the image's own transformers the same config loads fine. `:latest` was also re-pulled during attempt 4 and no earlier run, so the toolchain moved mid-diagnosis. Two things separated "malformed upload" from "moved toolchain": reproducing the real failing call (a bare AutoConfig load does not reproduce it; the trigger is reached through AutoTokenizer) and keeping zerofata's canonical tree, quantized cleanly on 2026-08-21, as a positive control. The fix drops `per_layer_config` rather than forcing global access. It is exactly redundant -- keys are precisely the ten full_attention layer indices, sole value (512, 4), verbatim the global fields -- and forcing instead would make `config.head_dim` answer 256 to the callers building the 512-wide layers. patch_perlayer.py re-proves that redundancy at apply time and refuses if it ever stops holding. Verified on the tensor table rather than the exit code: the output is identical family-for-family and count-for-count to the August canonical quant, with 356 BF16 vision-tower tensors preserved and input_activations=None. A GPU-free load leaves 0 tensors on meta and generates coherent prose. The section 4.4 serve test has NOT run -- GPU1 has 19.9 GB free against 19.5 GB of weights, so it needs a live seat displaced. Also fixes the A4B output, which had a truncation cap baked into its tokenizer (max_length 8192) from being quantized with the calibration corpus. Playbook gains section 3.17 for the pinned-transformers class and sharpens 3.16 to say drop the dataset outright for any A16 scheme. |
||
|
vh
|
eb75713c1b | Wire Beszel fleet filesystems, GPU telemetry, dashboard and alerts | ||
|
vh
|
6972e7ef7f |
feat(erp-seat): run 7 quantized to NVFP4A16 and serving as the trial seat on ana-ml2
- services/erp-seat-quant/run_quant_erp_v7.sh: v6 runner retargeted; dry-run gate passed identically (11,725 targets, 11,520 experts = 30x128x3, routers+vision BF16) - 49 GiB bf16 relayed gx10 -> ana-ml2 (no key path either way; nh3-dev relays), checksums verified against source; quant 49 -> 16 GiB, all post-steps clean - stacks/erp-seat: .env-driven swap to erp-tune-v7-nvfp4a16, served under its TRUE name; homepage labels + README updated, v6 rollback path recorded - stacks/litellm: trial -> erp-tune-v7-nvfp4a16 (config-file alias; /model/update refuses a config model, so this is an edit + restart) |
||
|
vh
|
c335c38c19 |
fix(booth): a partial ask answer is recorded, not refused
Operator: the form failed when a question was left blank. Refusing the whole submission over one blank threw away the picks that were made, and the HTML `required` on the radios blocked it in the browser before the server saw it. - answered questions recorded; blank ones land in `unanswered`; `complete` says whether the set is finished; a blank question carrying a note keeps the note - `required` dropped from both templates so the browser cannot block a partial - refused only when there is no pick anywhere AND no notes (a 400 — that would flip an open ask to answered with no decision recorded); a choice outside the option list is still an error - new ◐ partial state with an n/N count; skipped questions render as skipped - README + global CLAUDE.md tell reading sessions to check `complete` - 154 tests; v0.1.15 |
||
|
vh
|
784c555dbd | docs(booth): record WHEN inline ask placement earns its markup — the artifact-in-working-memory test, not just the placement rule (framing from tts-dev) | ||
|
vh
|
6e0b85ba27 | docs: refresh what today's work made stale — booth asks (inline placement promoted to its own section), ana-ml2 nvme7 settled by the scrub result, nh3-dev booth entry + the CLI-on-PATH fix, run-07 runbook outcome + serving state | ||
|
vh
|
78c3a7c170 |
feat(booth): asks render INLINE in a verbatim report, placed by the author
Operator verdict on the separate /asks page: the question belongs with the artifact it is about. A four-voice audition wants each voice's radio group under that voice's audio, and one submit for the lot. - booth/inline.py: data-booth-ask="stem" | "stem:key" | data-booth-ask-submit, plus <!-- booth:ask ... --> comments; unknown stem left alone, not blanked - _ask_inline.html: self-contained fragments (own scoped styles, no JS), per-question groups bound to one form via the HTML5 form= attribute so a scattered multi-question ask still POSTs once - unplaced questions and a missing submit block are appended, so a partially marked-up page can never produce an unsubmittable 400 - chip becomes a jump link to the first open ask; /asks page kept as a fallback - 6 tests (one caught the partial-placement drop); v0.1.14 |
||
|
vh
|
a56743ade3 |
fix(booth): asks were invisible in a booth serving its own index.html
A custom index.html is returned verbatim, so booth.html's asks panel never rendered there — a valid ask (emmie-anchor/anchor.ask.json) was listed by the CLI and shown nowhere, with nothing to say so. - panel extracted to _asks.html; new GET /b/<name>/asks standalone page - verbatim pages get an amber '? N open asks' chip beside the back chip - POST /answer honours back=asks so answering returns to that page - single-question asks now keep an optional 'title' (was silently dropped) - README + routes table; 8 regression tests; v0.1.12 |
||
|
vh
|
2a186e4762 | fix(booth): CLI resolves its source through the ~/.local/bin symlink (readlink -f), so ask/asks/answer/unlink work from any cwd | ||
|
vh
|
c85a700141 |
feat(booth): multi-question asks — a questions list renders one form with a radio group per question and lands as one answer sidecar keyed by question
- asks.py: single {prompt, options} and multi {title, questions:[{key, prompt, options, notes?}]} both normalise to questions[]; per-question notes; every question required on submit
- /answer reads choice.<key> / notes.<key> / notes for multi; single shape unchanged
- booth asks prints per-question picks; README + CLI header; install step symlinks the CLI to ~/.local/bin; v0.1.10; 135 tests
|
||
|
vh
|
3fe01225a9 |
feat(booth): asks — a multiple-choice question a session poses in a booth, answered by the operator as a radio form + notes, written back as an answer sidecar
- booth/asks.py (stdlib): <stem>.ask.json question / <stem>.answer.json answer; normalise+validate, atomic write, list with answer folded in, broken asks surfaced not hidden - POST /b/<name>/answer: validates choice against the ask (400), unknown stem 404, re-answer overwrites - booth.html asks panel above the gallery; amber open / green answered; JS-off form POST; index card + booth header badge for open asks - CLI: booth ask / asks / answer [--wait [SECS]]; remote sessions poll <stem>.answer.json over HTTP - ask/answer files excluded from gallery items and item counts; 23 tests; v0.1.9 |
||
|
vh
|
911ff20356 |
feat(erp-seat): NVFP4A16 quant pipeline for the Gemma-4 26B-A4B MoE ERP tune + ana-ml2 GPU1 serve stack
- services/erp-seat-quant/quant_nvfp4a16_gemma4_moe.py: linearize_moe first (playbook §3.15), asserts the expert Linear count, routers/vision/audio/norms/lm_head ignored, W4A16 for RP long-session fidelity, post-steps restore processor configs + template and reset the tokenizer truncation cap (§3.14); --dry-run proves targets before GPU time - services/erp-seat-quant/run_quant_erp_v6.sh: detached container on GPU1 (vllm-llmcompressor) - stacks/erp-seat: serve recipe copied from gemma4-charrp, true served name only, port 8021 |
||
|
vh
|
76fdf45925 |
feat(booth): pin/favorite, multi-select delete, newest-first link board
The standing link board grew from a flat oldest-first list with a per-row × into a manageable board: newest links lead, favorites stay on top, and several dead links can go in one pass. - Ordering: order_for_display() renders pinned rows first, then newest-first within each group (the board is an append log, so newest = most recently posted — the row you usually came to grab). - Pin/favorite: a per-row ★ toggles pinned state via POST /b/<name>/pin. State lives in a .pins sidecar dotfile (one content id per line), NOT inline in links.md — so links.md stays a pure atomic-append log (many sessions post concurrently) and a row's content id never changes just because it was pinned. remove_link_entry drops a removed row's pin; orphaned pins are inert (renderer only stars a live id). - Multi-select delete: checkboxes feed POST /b/<name>/unlink-many (repeated 'sel' content ids), with a select-all box and a live count. The per-row × stays for single removal. - One <form> with formaction buttons, so checkboxes, ×, ★, and bulk delete coexist without nested forms AND all work with JS off; JS only adds select-all and the live count. Per-row × confirm reads desc/url from data-* attrs, so an arbitrary posted description can't break into the JS. - Every action is keyed by content id, never row position — same race-safety the existing × has, extended to the bulk path. - Fixed pre-existing undefined --fg/--bg CSS refs in the board styles. Tests: +19 (pins round-trip, ordering, orphan-inert, remove-unpins, /pin and /unlink-many endpoints, board render + order). Full suite 102 passing. Deployed to nh3-dev booth.service; verified live (newest-first, pin round-trip, bulk delete) against the real 31-row board with no data loss. |
||
|
vh
|
0f748ea54e |
feat(searxng): move to nh3-docker, update, and expose as an MCP tool
The ana-docker instance was returning zero results for every query while reporting healthy — 4.5 months stale (2026.4.17 against a current 2026.9.3), its engine scrapers rotted against sites that had changed. /healthz proves the web app answers and says nothing about whether search works, so seven days of green sat on top of a search box that found nothing. Moved to nh3-docker rather than updated in place, because the colo egress is the other half of the problem: 38.120.12.42 is a datacenter address that DuckDuckGo and Startpage CAPTCHA, while nh3-docker egresses residentially at 70.230.226.88. Same reasoning as the fleet's residential proxy for yt-dlp, applied at the source instead of around it. Config corrected along the way: base_url said searxng.pfi.local, a name retired on 2026-08-19, while the environment said something else — the env won so nothing broke and the file quietly lied. The karmasearch.videos removal key never matched, because the engine's real name has a space. scripts/searxng-health.sh asserts results > 0 across three unrelated queries. That is the check that would have caught this, and the only kind that can: the mechanism was healthy throughout. services/searxng-mcp exposes it as `web_search` at user scope, so every Claude Code session has it. Zero results raise rather than returning an empty list — an empty list is indistinguishable from a broken aggregator, which is precisely how this hid. Old instance stopped and removed; DNS alias repointed to searxng.nh3.internal. |
||
|
vh
|
0ad332bb4a |
feat(booth): per-row link removal + render the link board as real UI
The standing link board is the one MULTI-WRITER booth -- every agent session appends operator-facing URLs to it. "Delete the folder" was the only granularity available, so removing one dead link meant hand-editing markdown. It is 32 rows and only grows. booth links row number, entry id, raw row booth unlink 3 by row number booth unlink 8b40e0a5 by entry id (what the UI's x posts) POST /b/<name>/unlink form field `entry` = content id ROWS ARE ADDRESSED BY CONTENT ID, NEVER BY POSITION. The board is append-only and multi-writer: another session can post between listing it and clicking x, and an index would then delete a neighbour. An id either matches the row you saw or matches nothing. A row number typed at the CLI is resolved to its id BEFORE anything is deleted. Appends and prunes now take the same flock on .links.lock, so a post cannot be lost inside a prune's read-modify-write. UI: a booth carrying links.md renders as rows -- description, URL, provenance, copy button, per-row x -- instead of a markdown blob. links.md is filtered out of the gallery so it does not appear twice; the header counts LINKS not files; the empty-state and the one-click "Wipe now" both stand down for a board (same rule as the kept lane: nothing durable is one click from gone). booth/links.py extracted, STDLIB ONLY. The CLI needs this logic and must not require the service venv -- importing app.py drags in FastAPI, so deleting a line from a text file would have needed a web framework installed. THREE BUGS FOUND BY TESTING, all in the shell wrapper while the module was correct throughout -- module-only tests would have caught none of them: - `[ "$n" -eq 0 ] && echo ...` as the LAST statement made `booth links` exit 1 whenever the board had rows. `unlink`'s index lookup calls it inside $( ) under `set -e`, so a successful listing killed the caller and the removal silently did nothing while reporting success. - ids are 8 hex chars and roughly one in forty is ALL DIGITS; those were read as row numbers, resolved to nothing, and removed nothing. Now disambiguated by the id's actual shape, not by "is it numeric". - filtering links.md out of the gallery left `items` empty, so a full board rendered "This booth is empty" and an empty <div class="gallery"> under 32 visible rows. 87 tests (was 76): parser tolerance of hand-written prose, content-id stability across concurrent appends, removal precision, UI branch behaviour for board/normal/empty booths, and subprocess CLI tests pinning the two shell bugs. Deployed to nh3-dev and verified against the live 32-row board read-only; board file byte-identical afterwards. |
||
|
vh
|
4be880f36c |
feat(booth): kept boards can be deleted from the UI; document the TTL-reset trap
Kept boards had no delete path in the UI at all. The kept lane deliberately omits the wipe control -- a one-click wipe next to the durable stuff is a footgun -- but "deliberate" had been implemented as "impossible": the only routes out were ssh or a hand-written API call. Now it is two deliberate acts. A `release` control on kept cards drops the sentinel, the board moves to the ephemeral lane, and the existing x wipes it from there. Release is reversible -- POST /b/<name>/keep pins it again. POST /b/<name>/unkeep release the pin POST /b/<name>/keep pin it (round-trip, so release is not a one-way door) FOUND WHILE TESTING, and it invalidates the previously-documented workaround: removing the sentinel BUMPS the booth directory's mtime, and booth age is the newest mtime in the tree -- so a released board's clock RESETS from 10,000s to 0s and it survives another full TTL. The old comment said "remove the sentinel first (it rejoins the sweep)", which is true but means the board lives another 24h, not that it gets reaped. Unkeep-and-wait is a delay, not a delete. test_releasing_a_board_RESETS_its_ttl_clock pins that behaviour deliberately so nobody re-derives the workaround. Release is what unlocks the x; the x is what deletes. CLI: `booth rm` already worked on kept boards but said nothing about it. It now announces "(was KEPT -- durable board)" so wiping something durable can never look identical to wiping run output. Not a block -- a CLI user naming a booth is being explicit. 5 new tests (67 pass). Verified live on nh3-dev: release renders on all four kept boards, the ephemeral lane keeps its x, and the links board is untouched with its sentinel intact. |
||
|
vh
|
0755ba7d00 |
fix(quant): stop baking the calibration truncation cap into the shipped tokenizer
load_calib tokenizes with tok(..., truncation=True, max_length=seqlen). For a
fast tokenizer that mutates the Rust backend's truncation state in place, and
the subsequent tok.save_pretrained() persisted it, so every mixed-NVFP4 build
shipped a tokenizer.json carrying
"truncation": {"direction": "Right", "max_length": 2048, ...}
against a source whose value is null. Every prompt was clamped at the
calibration length, permanently.
It hid because older transformers does not enforce the text-vs-ids count
check. On a newer one the seat dies at startup with a message that names
images and never mentions tokenizers:
ValueError: Mismatch in `image` token count between text and `input_ids`.
Got ids=[2047] and text=[16384].
The cap also silently limited image resolution well before it killed
anything -- at 2048 the largest servable image is about 1448x1448, since
(edge/patch)^2 / merge^2 image tokens have to fit under it.
Fix saves a pristine tokenizer re-read from the source rather than the
mutated calibration object, and then asserts truncation is null so the
defect fails the build instead of shipping again.
Playbook gains section 3.14 with the symptom, the cause, the audit one-liner
and a table of which builds were affected, plus a fourth mandatory post-step.
The transferable lesson is called out: this is the third case of an artifact
carrying config authored against an older transformers that a newer one
begins enforcing, so an image bump is a config-compatibility event rather
than just a version change.
|
||
|
vh
|
36c173c6a1 |
feat(mog-sec): quant + serve M.O.G.-SEC pen-test seat; PPL on gen; retire fable
Autonomous overnight run under the operator's full-autonomy grant. End state:
fleet up, gen seat untouched, a new verified pen-test seat serving where fable was.
PPL on the orcarouter gen seat (fable downed to free GPU1 for a nospec probe,
probe torn down after): mean 7.07 / median 5.76, within noise of heresy 6.910 /
5.625 and identical to our recipe's usual 7.059. The gen-seat search is settled.
M.O.G.-SEC: chose Blackfrost-Research/M.O.G.-SEC-27B-1M-CTX-BF16 (rev deede677)
over the pre-made ModelOpt NVFP4, which was disqualified on W4A4 4-bit activations
(the AEON degradation mode, catastrophic on a 1M-context model), zero MTP tensors,
and ModelOpt format. Pulled, format-screened (P(<think>) 1.11e-05, clean), quanted
in-house to mixed NVFP4+FP8 (23.4 GB, MTP + vision preserved), and served in the
retired fable slot.
stacks/mog-sec ana-ml2 GPU1 :8019, KV 418,218 tok / 1.60x @ 262K
aliases mog-sec (non-thinking), mog-sec-reasoning (thinking)
gates surface 6/6, MTP 55.3%, format 0/15 leak, vision 7/3/1,
capability 4/4 (delivers offensive-security content)
Served at native 262K, NOT the card's 1M -- the 1M needs YaRN (absent from the
weights' config) plus the SGLang/DFlash2 path the repo ships a deployment kit for,
neither of which is our vLLM surface. A real 1M seat is a separate SGLang project.
Retired char-rp-reasoning + char-rp-fable (zero traffic, pointed at the downed
fable :8019; now 404 cleanly, not repointed -- a security model is not an RP model).
char-rp (meromero) untouched. Vision preprocessor built from the model's own
image_processor block, same trick as the MeroMero seat.
GPU0 seats (gen, meromero) were untouched and healthy throughout. The quant ran in
GPU1 free space with no production seat stopped except fable, which was replaced.
|
||
|
vh
|
e4576f0989 |
test(gen-seat): PPL on orcarouter — mean 7.07 / median 5.76, within noise of heresy
Measured against a spec-decode-free probe on GPU1 (fable downed to free the VRAM, probe torn down after). eval_quality.py aborts PPL under --speculative-config, so a nospec probe is the only way to read it. orcarouter mean 7.0655 median 5.758 heresy mean 6.910 median 5.625 (+2.2% mean) our recipe's usual mixed-quant PPL: 7.059 -- orcarouter is identical to it So orcarouter matches heresy on fidelity and wins on every other axis: MTP acceptance 58.4% vs 47.2%, zero think-leak, vision 7/8. The gen-seat search that ran through Cold-Fusion, heresy, and preetpatel is settled on orcarouter. |
||
|
vh
|
ce09ac4fa6 |
test(gen-seat): add a real vision battery — orcarouter scores 7/8
surface_test.py's vision check is one image and one word. It proves the tower loads; it does not prove the tower works. This battery uses generated images with known ground truth so every answer is objectively gradeable. Against orcarouter NVFP4-mixed on the `gen` alias: T1 OCR, 5 lines incl. one at 18px PASS all 5 exact T2 counting + attribute binding PASS 7 circles / 3 triangles / 1 square T3 bar chart, 6 values + max/min PASS 6/6 exact T4b occlusion, star behind rectangle PASS T4c aspect ratio of a 160x140 rectangle FAIL called it taller than wide T5 two images, which has text PASS T6 four images, the seat's cap PASS all four named T7 five images, one over the cap PASS rejected with HTTP 400 No <think> leak on any vision call. The single miss is fine-grained relative-dimension estimation on a near-square shape, and it reproduced across two runs (the longer T4 called the same rectangle "equal width and height"). Counting, OCR, chart values and occlusion ordering are all solid, so this is a precise-geometry weakness, not a broken tower. Recorded so nobody builds a feature on this model judging relative sizes. T7 earns its place separately: it confirms the per-prompt image cap fails loudly with a 400 rather than silently dropping the extra image. |
||
|
vh
|
f85d102813 |
test(gen-seat): orcarouter passes every gate — in-band MTP head delivers +11 points
Gates run against the live seat while the operator tested in parallel. <think> leak (n=30, 4 prompt types + multi-turn) 0/30, 0 empty MTP acceptance 58.4% @ 117.11 tok/s median surface 6/6 abliteration survival 4/4 compliance deterministic quality gens coherent and correct PPL still blocked For scale on the leak gate, the abandoned h300 build scored 8/30 on this exact instrument, and its abliteration-survival samples had 2 of 4 open with "<think>Ok, let's figure this out:". Orcarouter has none. The headline is MTP acceptance. 58.4% against heresy's byte-identical base head at 47.2% is +11 points, and it sits level with our own in-band L35 at 59.1%. That is the additive in-band-vs-graft delta the entire Cold-Fusion experiment was built to measure and never cleanly delivered -- orcarouter handed it over for free because the author had already done the Robinson edit on the head. Surface 6/6 covers plain chat, vision, tool calling, the thinking split, a 36,042-token long-context retrieval, and streaming. PPL remains blocked on a spec-decode-free probe seat: it needs ~22 GB and GPU1 has ~16 GB free. Comparison target is heresy at 6.910. |
||
|
vh
|
c8f128bdff |
feat(gen-seat): quant orcarouter — its MTP head is already Robinson-abliterated in-band
Pulled orcarouter/Qwen3.8-27B-Uncensored at rev 9878936b (55.5 GB, gated, our token has access) and built /tank/aimodels/qwen38-27b-orcarouter-nvfp4-mixed (23.4 GB, mixed NVFP4+FP8). Verified, not yet cut over. The operator asked whether we could apply the Robinson path to the MTP head. We cannot, because the author already did. compare_mtp_head.py against the verbatim base graft: 13 of 15 tensors byte-identical, exactly 2 differ -- mtp.layers.0.self_attn.o_proj.weight and mtp.layers.0.mlp.down_proj.weight, which are precisely the two residual writers our own abliterate.py targets (EXPECT_MTP_WRITERS = 2). Reverse-engineered the edit from the weights alone (mtp_delta.py, added here): sigma2/sigma1 = 0.0164 on BOTH tensors rank-1, a single-direction projection |cos| between the two recovered dirs = 1.0000 ONE shared direction ||delta||/||W|| = 1.42% and 1.41% a gentle, consistent projection sink energy dim 3994 = 0.0000% sink-clean; Heretic's was 6.18% That is the Robinson in-band MTP abliteration, already applied, with a direction that passes our sink screen outright. Nothing to do but preserve it, and the quant carries it byte-identically. This is the configuration the entire Cold-Fusion experiment was designed to test and never cleanly delivered. The new format screen paid for itself on its first real use: think_prior.py on the bf16 BEFORE any GPU time gave P(<think>) = 1.23e-06 at rank 52, against Cold-Fusion stock 0.1850 and h300 0.2216. Roughly 150,000x cleaner. Two durable findings about the pipeline itself: The quant needs ~17 GB, not a whole card. It ran entirely in GPU1's spare 16 GB with ZERO production seats stopped -- the h300 run's "stop BOTH GPU0 seats" was never necessary, it simply had a free card by coincidence. The first attempt OOM'd by 2.37 GiB at layer 64 of 65 with 3.57 GiB reserved-but-unallocated, which is fragmentation, and PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True closed it. post_quant.py now builds a missing output index from the safetensors headers. A sub-23 GB quant saves one bare shard with no index, and post_quant needs one; this has broken three separate rounds and been hand-fixed every time. The header is read by struct-unpacking the u64 length and parsing the JSON -- never safe_open, which mmaps the whole 22 GB shard and ENOMEMs on ZFS. Artifact verified: mixed-precision, 1968 tensors, 15 mtp, 333 visual, re:^mtp.* present in the ignore list (llm-compressor pruned it as always), preproc restored. Imatrix deferred per operator; the log confirms the usual uniform-MSE fallback, so this build stays apples-to-apples with heresy's PPL 6.910. |
||
|
vh
|
37e9e1ca7f |
revert(gen-seat): abandon Cold-Fusion, roll back to heresy — the leak is in the base
Operator directive, given before the result was in: if it's the base, abandon h300 and the base too. The dose-response said base (18.5% of 22.2%), so it fired. Live gen seat is /tank/aimodels/qwen38-27b-heresy-nvfp4-mixed again, restored from .env.bak-coldfusion-L35-20260820. The h300 env is preserved at .env.bak-h300-abandoned-20260821. The clincher, same probe pointed at heresy: Cold-Fusion stock P(<think>) 0.1850 Cold-Fusion L35 0.2048 Cold-Fusion h300 0.2216 heresy (restored) not in the top 20, <0.002 A >100x gap between the families, which is why no rollback inside Cold-Fusion would have helped -- stock and L35 leak at nearly the h300 rate. Verified after rollback: 0/30 leaks and 0 empty on the same instrument that scored h300 at 8/30, with the EXISTING enable_thinking:false config; KV pool 403,065 tok / 1.54x, heresy's exact documented baseline; 7/7 aliases; vision intact. No LiteLLM change was needed, so the chat_template_kwargs fix is left unapplied -- it worked, but it was a workaround for a base we no longer serve. Cost, stated plainly: 8/100 refusals becomes 29/100, a 3.6x regression on the axis the whole Heretic-300 run existed to move. Accepted deliberately. What carries forward is the methodology, none of which lived in the Cold-Fusion weights: direction_scope=0 beating per-layer on a merged base, aggression not being the lever, PR #317 silently dropping the MTP head on save, the MPOA and sink-screen reasoning, the graft/KL/catatonia/export harnesses, and the finding that a pristine MTP graft accepts as well as an in-band edit. New acceptance gate earned here: run think_prior.py on a candidate's STOCK weights before committing GPU time. It is a ~10s CPU measurement and it would have disqualified Cold-Fusion before the 300-trial study ever started. Heretic's objective has no format-compliance term at all -- the same blindness that removed the self-harm guardrail. Nothing deleted. Every Cold-Fusion artifact, the 300-trial Optuna journal and catatonia-T260.json remain on disk. Abandon means stop serving, not rm. |
||
|
vh
|
5ee2325820 |
feat(coldfusion-abliteration): dose-response settles the <think> leak — base 83%, our abliteration 17%
Answers "how likely is it that our abliteration caused this?" with a measurement instead of a prior. P(<think>) at the first generated token, template rendered enable_thinking=false so the prompt already carries a CLOSED think pair -- the exact event behind the leak. Raw softmax, bf16, CPU-only, one process per model. Deterministic: stock reproduced to 17 significant figures across two runs. coldfusion-bf16 none (stock) 0.1850 rank 3 coldfusion-abliterated-L35-bf16 Robinson L35, mild 0.2048 rank 2 coldfusion-h300-mtp-bf16 Heretic-300, heavy 0.2216 rank 2 The stock, untouched base already puts 18.5% of first-token mass on opening a think block the template had closed. Abliteration adds a real, monotonic, dose-dependent +3.7 points -- a nudge on a pre-existing base, not the cause. Cold-Fusion is a reasoning-token-compression finetune, i.e. a model trained to think briefly, and the leak's text shape agrees: a compact correct trace with a trained transition marker, which is trained behavior rather than damage. This changes the options. Rolling back to L35 or stock does NOT fix the leak -- at 18.5% under temp 0.7 / top_p 0.8 they leak at nearly the h300 rate. Only leaving the Cold-Fusion family escapes it, at the cost of the 8/100 refusal result. The chat_template_kwargs fix is the correct lever. Durable methodology point: a forward-KL budget cannot catch this. Heretic minimizes forward KL(stock||abliterated), which is near-blind to the model putting new mass on tokens stock barely used -- that is reverse KL's job, and we measured exactly that asymmetry on L35 (reverse 1.43 vs forward 0.70). h300's KL of 0.0136 is not evidence of innocence. For any "did the abliteration break behavior X" question, measure P(token) directly. Ran CPU-only deliberately: 96 EPYC cores and 265 GB of RAM make a 27B forward pass cheap, so this cost no GPU window and no seat downtime, where the obvious route was stopping both GPU0 seats. Also normalizes two more abliteration output dirs from root-owned 0600 to llmuser 0664. The unreadable-model failure surfaces as FileNotFoundError rather than a permission error, which is worth knowing before it wastes a run. |
||
|
vh
|
91f4cf22e1 |
fix(gen-seat): diagnose the unterminated-<think> leak — model defect, temp-triggered
Operator reported the new Heretic-300 gen seat "sends CoT but never completes
the turn" through Lobe. Diagnosed; not yet fixed (the fix changes gen's
semantics, so it is the operator's call).
The Qwen3.8 chat template appends a pre-closed <think>\n\n</think>\n\n when
enable_thinking is false. The h300 model opens a fresh <think> anyway and never
closes it. Because the prompt already closed the block, vLLM's qwen3 reasoning
parser is not in reasoning state, so the tag passes through as ordinary text --
reasoning_content empty, reasoning_tokens 0, and the whole reasoning-plus-answer
blob lands in content. Lobe then correctly treats the unterminated tag as
still-thinking and renders no answer. The client and the serving stack are both
behaving correctly; the model is not.
The trigger is TEMPERATURE, not presence_penalty (n=12 per arm):
temp 0.7, pp 1.5 (current gen) 4/12
temp 0.7, pp 0.0 4/12
temp 0.7, pp 0.5 3/12
temp 0, pp 1.5 0/12
That falsifies the standing hypothesis, recorded in the litellm config comment
and in the operator's own 2026-08-16 note, that presence_penalty 1.5 is the
first dial to move. It is not this bug's cause.
It also explains the blast radius: only the two temp-0.7 aliases leak, `gen`
and `summarizer-large`. summarizer, classifier, image-judge and qwen-image-bench
all run at temp 0 and are clean, so nevermore's summarizer path is unaffected.
Candidate fix, validated n=30 over 4 prompt types plus a 3-turn conversation:
chat_template_kwargs {enable_thinking: true, reasoning_effort: low} takes 8/30
leaks to 0/30, at ~+27% completion tokens and a ~3% empty-content residual.
The tell appears in eval_coldfusion_h300.json and in none of the aeon, heresy,
mixed or w4a16 evals, so it is new with this build -- but L35 was never evaled,
so this does not separate a Cold-Fusion base trait from a Heretic-300
abliteration artifact.
Reproducers and the full method land in bench/think-leak/. Note in particular
that the 7/7 alias smoke test run at cutover structurally could not catch this:
trivial prompts never invite reasoning, so they never sample the leaking token.
|
||
|
vh
|
e3ce713f7f |
feat(gen-seat): cut over to Heretic-300 — 7/7 aliases, vision intact, MTP 59.7%
Live GEN_MODEL is now qwen38-27b-coldfusion-h300-nvfp4-mixed (ana-ml2 GPU0 :8015). Served-name left unchanged so all 7 LiteLLM aliases route without a gateway edit. Verification: KV pool 401,550 tok / 1.53x (baseline 403k / 1.54x) LiteLLM aliases 7/7 green vision 3/3 shapes, colour+form+position correct MTP acceptance 59.7% median @ 118.37 tok/s quality gens 4/4 correct abliteration 4/4 compliance PPL NOT measured (see below) The roadmap predicted ~47% acceptance for a pristine MTP graft versus L35's 59.1% in-band edit. Measured 59.7% on the same harness: there is no acceptance penalty, which removes the throughput argument for reimplementing MPOA. A single long-prose generation read 47.5% off the same counters -- below the 8-run minimum of 49.0% -- and would have "confirmed" the prediction by coincidence. Acceptance must be read from quickbench.py, never one sample. PPL is blocked on VRAM, not on the model: eval_quality.py aborts with "prompt_logprobs look uniform" under --speculative-config, and the probe-seat workaround needs ~22 GB while both cards sit at ~96% committed. Also normalizes the quant dir from root:0600 to llmuser:llmuser 0664 to match every other model dir, and records that config.json sha256 is byte-identical across the h300 and L35 quants and is therefore useless for confirming which weights are mounted (mtime and a head-hash are the discriminating views). Rollback is one line to .env.bak-pre-h300-20260820. |
||
|
vh
|
f90a5025de |
feat(coldfusion-abliteration): Heretic-300 — 8/100 refusals at KL 0.0136, beats the heresy bar 3.6x
Ran Heretic v1.4.0's 300-trial TPE search on Cold-Fusion. Best trial scores 8/100 refusals at KL 0.0136 against a 98/100 base, versus absolute-heresy at 29/100 and our hand-tuned Robinson L35 at 72/100 / KL 0.0116 — i.e. 64 fewer refusals for the same damage. Hand-verified coherent: correct arithmetic with shown working, clean code, 66-167 word prose across nine probes. Durable findings: - direction_scope=0 (single shared direction) is decisive on this merged base: n=129, best 8/100. Per-layer directions n=131 never beat 52/100 despite a better median. Points against the multi-direction intuition for a diffuse direction (our two-template |cos| is 0.62 vs Robinson's 0.99 on stock). - Aggression is not the lever. r(KL, refusals) = -0.561 over 261 trials; the KL<0.02 band contains both the worst results (median 87/100) and the single best. A KL 0.3554 trial scored worse than one at 0.0193. - PR #317 confirmed: Heretic silently drops the MTP head on save. Source 1199 tensors -> export 1184, all 15 mtp.* gone, vision 333/333 intact, exit 0, no warning. This is also why absolute-heresy ships a byte-identical MTP head — a bug, not a design choice. Always diff tensor keys after a Heretic export. - Heretic's recovered direction carries 6.18% of its energy in sink dim 3994, versus 0.094% for our L35 and 1.97% for the L39 we rejected as brick-inducing. It survives that only because of magnitude-preserving ablation (row_normalization=FULL); our plain projection has no such protection, so the sink screen correctly refused the in-band MTP graft. Same direction, different operation. MPOA is the prerequisite for in-band MTP on a Heretic trunk. - Heretic's edit is recoverable from weights: delta is rank-1 (s2/s1 ~ 0.010), SVD gives the direction, norms give per-layer weights (1.08 -> 1.34, i.e. over-projection). Cross-layer |cos| agreement 0.9903 independently confirms the single-direction result. New tooling in services/coldfusion-abliteration/: kl_divergence.py first-token KL, class-split, zero noise floor catatonia_gate.py 12 probes x 220 tokens, prints every completion heretic_export.py PTY driver; selects by measured value, never by menu position — Heretic's resume prompt puts "delete the checkpoint and all results" one arrow-key from the target graft_mtp.py recovers the trunk direction by SVD; --pristine for the safe path when the sink screen refuses Also adds quant playbook 3.13: the NVFP4 recipe sets observer="imatrix_mse" but llm-compressor has always silently fallen back to uniform MSE for want of importance data — on this build and on the incumbent. Existing A/B comparisons stay valid since every build shares the fallback. Parked as id 42. Guardrail note: this build has lost the self-harm guardrail that the Robinson L35 build retained. Restoration is the operator's own work item. |
||
|
vh
|
a9d73dad41 |
fix(coldfusion-abliteration): GPU0 seat restore order is load-bearing — correct the claim and the runbook
Restoring the two GPU0 seats with `start meromero; sleep 10; start gen` put
meromero into a 7-restart crash-loop:
ValueError: Free memory on device cuda:0 (35.3/94.97 GiB) on startup is less
than desired GPU memory utilization (0.52, 49.38 GiB).
The previous commit's README claimed restore order "is not actually load-bearing"
on the grounds that both seats pass --gpu-memory-utilization as a fraction of
total VRAM. That is half right and the wrong half mattered: the fraction sets the
target, but vLLM gates startup on FREE VRAM and refuses to start unless the whole
target is available. GPU0 runs at ~96.4/97.9 GB with roughly 0.4 GiB of slack, so
the seats coexist only in the order they were originally brought up, and meromero
is the one that does not fit in the remainder. The pre-existing auto-memory note
("gen takes a fraction of free VRAM at startup and will starve meromero") was
pointing at the real effect.
Also: "first" means healthy, not ten seconds earlier. A sleep 10 against a
two-to-three minute weight load is simultaneity, not ordering — gate on observed
state.
Recovery applied: stop gen, wait for meromero healthy, start gen. Verified
against the pre-window baseline rather than against "both green":
gen KV 14.36 GiB / 403,065 tok / 1.54x -> 14.34 GiB / 401,550 tok / 1.53x
meromero KV 542,202 tok -> 542,202 tok
RestartCount 0 on both; summarizer smoke-tested through LiteLLM
Note for the next reader: raw nvidia-smi used-MiB is the wrong check here. It
reads 89,503 now vs 96,376 before, which looks like a 6.9 GB regression and is
allocator slack — serving capacity is unchanged. The anomalous boots were the
high ones (34.95 GiB KV), where gen came up on an empty card mid-window.
|
||
|
vh
|
1b3fb270e7 |
feat(coldfusion-abliteration): first-token KL measured — 28.4x selectivity, harmless median 0.0211
Adds `kl_divergence.py`: first-token KL(stock || abliterated) over the full 248,320-token vocabulary, bf16 vs bf16, scored separately for held-out harmless and reserved-harmful prompts. Result (L35, 256 harmless / 104 harmful, answer mode): harmless median 0.0211 mean 0.0364 top-1 agreement 89.8% harmful median 0.5996 mean 0.6992 top-1 agreement 55.8% selectivity 28.4x (72.8x in think mode) Self-KL noise floor is exactly 0.0, and all 720 per-prompt values are bit-identical between a single-process and a two-process run, so the figures are signal rather than bf16 jitter. Reverse KL on harmful/answer is 1.43 vs forward 0.70 — the mass-where-stock-had-none asymmetry expected of a refusal-direction removal. Against the Heretic reference figures (0.1191 prior seat, 0.0759 the live absolute-heresy seat) this is materially gentler, but those are the other tool's optimizer output on a different base with its own harmless set and template — order-of-magnitude, not head-to-head. KL remains a fidelity number; the viability gate is still MTP acceptance (59.1%). Method notes: - Prompt classes are reported separately by design. A single averaged KL over a mixed corpus is close to meaningless, since the metric is meant to be large on harmful prompts and small on benign ones; the ratio carries the information. - The harmless evaluation set is drawn from the alpaca pool minus calibration's own draw, reconstructed by replaying that draw rather than remembered, and asserted disjoint on text. The harmful set is the reserved test split. - `render` is imported from abliterate.py rather than copied, so the measurement cannot drift from the rendering the direction was captured against. - Batch size 1 with logits_to_keep=1: no padding semantics, ~0.6 MB of logits. Three corrections to the runbook, each of which cost time: - "bf16 is 50 GB, only gen must go" was 50.10 GiB mislabelled. Text-only weights are 51,300 MiB; freeing either GPU0 seat alone leaves ~50,933 MiB. Both must stop. VRAM is now sized from the safetensors headers at run time. - A 27B model cannot be released in-process: `del` + gc + empty_cache left free VRAM at 45,287 MiB, and so did confining the model to an inner frame that exits. Only process exit returned the card (96,689 MiB). The first run completed only because the allocator hit OOM, collected, and retried. Each model now gets its own process, handing log-probs to disk between stages. - The residency gate read hf_device_map, which transformers leaves empty when the model fits on one device — it reported "(unsharded)" whether or not anything was wrong, so it could never fail. It now reads parameter devices directly. Model-agnostic lessons promoted to the quant playbook (new 3.12). |
||
|
vh
|
725c8fdf9e |
feat(coldfusion-abliteration): THESIS PROVEN — in-band-abliterated MTP head accepts 59.1% (beats incumbent ~47%)
Quantized the L35 abliterated model to mixed NVFP4 and measured MTP acceptance
end to end. The experiment's whole premise: Heretic (the incumbent gen seat)
leaves the MTP head a byte-identical base graft its wrapper never loads, whereas
Robinson abliterates the MTP head in-band — the question was whether that in-band
edit survives well enough to spec-decode. It does, better than the graft:
MTP acceptance 59.1% median (51-65%, 8 cache-busted topics) vs incumbent ~47%
decode 118.7 tok/s median (faster; image-confounded, read as not-worse)
abliteration survives quant (creative refusals drop, self-harm guardrail
intact, coherent)
Output at /tank/aimodels/qwen38-27b-coldfusion-L35-nvfp4-mixed (22.5 GB). Result
JSON in bench/. NOT cut over — the incumbent seat is untouched; making L35 the gen
seat is a separate decision needing the full Stage-3 gate + real multi-turn hold.
Two env foot-guns hardened along the way:
- quant_mixed_nvfp4.py now promotes text_config attention fields
(num_attention_heads etc.) to the top-level config for the oneshot, then
restores. transformers 5.10 / llmcompressor 0.12 (this venv moved under us
since the Aug-15 heresy quant) no longer delegate the top-level lookup, so
oneshot raised "Cannot determine num_attention_heads". Same "the fight is the
environment" pattern as the abliteration capture.
- a sub-~23GB quant saves as a single model.safetensors with no index, so the
post_quant MTP graft needed an index built first — from the safetensors header,
not safe_open (which mmaps the whole shard and ENOMEMs on ZFS).
post_quant grafted the abliterated MTP (15 tensors, 849 MB) and re-injected
re:^mtp.* into quantization_config.ignore (llm-compressor pruned it again — the
two-rounds-lost 0%-MTP bug, fired and repaired as designed). Probe served on the
pinned nightly (#51113 qwen3_5_mtp fix) to match the live seat's vLLM.
|
||
|
vh
|
e9dbc8660b |
feat(coldfusion-abliteration): abliteration LANDS at layer 35 — separation selector, shard-surgery write, three false diagnoses corrected
The abliterated model works. A/B vs stock on a matched greedy battery: explicit sexual + graphic torture (the measured stock refusal surface) go from refused to complied/engaged, held-out AdvBench prompts loosen, the self-harm guardrail survives, coherence intact — the Robinson design point exactly. Output at /tank/aimodels/qwen38-27b-coldfusion-abliterated-L35-bf16, verified bitwise: 131/131 targets changed, 333/333 vision byte-identical (delta 0.0), 735/735 others untouched. Getting there corrected three diagnoses the prior session had backwards. 1. The layer-selection metric was wrong, and that was the whole ballgame. The recipe picks the abliteration layer by peak two-template |cos| agreement. On this heavily-merged base that metric is anti-correlated with efficacy: its argmax (layer 18) is the WORST-separating layer in the window (Cohen's d 5.51 vs 9.89 at the peak), and abliterating there was a measured behavioral no-op — stock and "abliterated" refused all six probes identically. Cause: the two renderings end in different generative modes (</think> vs <think>), so |cos| scores answer-vs-reason mode, not refusal, and on a merge the mode term dominates. Replaced selection with harmful/harmless SEPARATION (Cohen's d / AUC of the direction's projection), gated on the sink screen since separation and sink-energy both climb with depth. Picks layer 35 (d 9.35, AUC 0.9997, sink 0.094%). Agreement is kept as a printed diagnostic. 2. The "bf16 NaNs, use fp32" rule was a misdiagnosis. The NaN was never precision — it was multi-GPU sharding (the residual stream zeroes two layers past the GPU0->GPU1 boundary; the first capture's layer 22 happened to sit in the healthy region, which is why it looked fine) plus PYTORCH_CUDA_ALLOC_CONF=expandable_segments (corrupts retained tensors; the corruption MOVED between bit-identical forwards, the tell that it was memory not math). On one GPU with a plain allocator, bf16 full-64-layer is exactly deterministic and coherent, at 50 GB and 4.3x the throughput of the 111 GB fp32 it replaced. Both defects are now hard gates (residency exit 8, allocator exit 9); capture pins CUDA_VISIBLE_DEVICES=0. 3. The corpus-size hypothesis was falsified. 52x more calibration data (8->416, mlabonne/harmful_behaviors = the recipe's actual AdvBench split, already on the box) moved agreement 0.594->0.624 — nothing. Kept the 416/416 corpus anyway (calibration.py); it gives the clean separation signal. The held-out 104-prompt test split is reserved and asserted disjoint. Also: the --out write is now shard-level surgery (reads/writes the 18 safetensors directly, no model object, no GPU). This is correctness, not thrift — AutoModelForCausalLM resolves to the TEXT model, so save_pretrained would drop all 333 vision tensors AND skip the MTP head (the in-band MTP edit is the entire point of the Robinson formula). Neither failure raises. Shard surgery makes vision and the other 1068 tensors byte-identical by construction. Batched capture with a dtype-aware equivalence gate; hidden states captured via forward pre-hook (reading output_hidden_states off the returned object is unsafe here — buffers get recycled). Sharding/allocator lessons promoted to the quantization playbook (model-agnostic, sections 3.9-3.11 + superseded table); the selection-metric lesson added to the recipe doc. The dead layer-18 no-op checkpoint was removed (52 GB, confirmed identical to stock). Incumbent gen seat untouched. Full canonical refusal-probe re-profile and MTP-acceptance-on-quant still owed before this becomes a gen-seat candidate. |
||
|
vh
|
f714f28195 |
feat(coldfusion-abliteration): Robinson's real 416-prompt corpus, batched capture, two new gates
The 8/8 calibration set gave |cos| agreement 0.594 against the recipe's 0.9925. This wires in the corpus the recipe actually used and makes a capture at that scale affordable. Corpus (calibration.py, new). The recipe's "held-out train/test split of 416/104 with overlap 0" names mlabonne/harmful_behaviors exactly — 416 train / 104 test, AdvBench-derived — and it plus harmless_alpaca were already staged in ana-ml2's HF dataset cache. Read via pyarrow, no datasets dependency, no hub access. Harmful is order-deterministic (no seed), so a re-capture is reproducible from the flags alone. The 104-prompt test split is reserved as the held-out generalization probe and asserted disjoint, so the post-write re-profile cannot silently become in-distribution. --calib builtin reproduces the legacy run. Batched capture. 832 prompts x 2 templates = 1664 forwards. Padding is on the RIGHT: in a causal stack nothing after position t reaches position t, so trailing pads cannot touch the token read, whereas left padding feeds pads into the DeltaNet recurrence ahead of the prompt — the path whose torch fallback already NaN'd once here. Means accumulate in float64; the direction is a difference of means, which is where cancellation lives on this model. Gates added, both protecting numbers rather than tensors: - batch-equivalence: proves padded-batch == single-prompt (rel 1e-3) before spending the capture window. - surgery pre-check: aborts if any of the 131 targets is absent or on the meta device. orthogonalize_ edits in place, and an in-place write to an accelerate-offloaded tensor is a silent no-op — that ships a half-abliterated model past a smoke test. Fixed a reporting bug: the agreement line printed the global agree.max() beside the window's argmax layer, so the first capture read as 0.8538 when the real in-window number was 0.5944. The global peak sits in the early layers where the dim-3994 massive activation inflates agreement for reasons unrelated to refusal. Now prints window max, a top-5, and labels the global figure informational. --max-layer truncates the decoder for capture. Exact, not approximate: a causal stack's layer-N state cannot depend on layers above N, so any value above the window top leaves the direction bit-identical while cutting fp32 residency and forward cost. 46 drops 18 of 64 layers and is what keeps fp32 off CPU offload. Refused on the write path, where it would emit a truncated checkpoint. Verified on ana-ml2 without the GPU: dry-run still 1:1 (131 tensors, all coverage gates), calibration loads 416/416 deterministically with its guards firing, both --max-layer guards exit as designed. Also confirmed against chat_template.jinja that enable_thinking=True does resolve reasoning_effort to xhigh, so the two renderings are the recipe's — template selection was not the cause of the low agreement. The re-capture itself is unrun: it needs the fp32 VRAM window and therefore production seat downtime. |
||
|
vh
|
7abd3011f7 |
fix(coldfusion-abliteration): capture works — fp32 forward + finite-gate
The --capture forward NaN'd repeatedly. Root cause: transformers' Qwen3.5 DeltaNet linear-attention needs the causal-conv1d fast-path kernel, which can't be built here (no nvcc, no prebuilt wheel). Its torch fallback produces nondeterministic all-NaN hidden states in bf16 -- same 11-token input finite on one forward, NaN at layer 4 on the next. bf16 and fp32 share exponent range, so it's precision-driven catastrophic cancellation, not overflow, and fp32 resolves it. Fixes: - --capture now loads fp32 (the write/surgery path stays bf16 -- no forward, no NaN). attn_implementation=sdpa pinned. - A finite-gate aborts on a non-finite direction. The sink screen alone can't catch this: nan > threshold is False, so a NaN direction "passed" it and got saved silently on the first run. Capture result (fp32, full GPU): refusal direction finite, unit-normed, layer 22, sink energy 0.0008% in dim 3994 -- clean, not sink-dominated. Saved. Caveat recorded: two-template |cos| agreement is 0.59 at layer 22 vs Robinson's 0.99, almost certainly the small 8/8 calibration set vs their 416/104. Valid but noisier than ideal; the README flags expanding the sets before the write. README documents the three environment gotchas (fp32-for-capture, the seats that must be stopped for the 110GB fp32 VRAM and how to restore them, and the fla side-dir PYTHONPATH) so the next run doesn't rediscover them. |
||
|
vh
|
b56cb0db13 |
docs(coldfusion-abliteration): dry-run passed — recipe maps 1:1 (131 tensors)
Dry-run against the fully-staged bf16 confirms the Robinson recipe transfers onto the DavidAU Cold-Fusion checkpoint with no name drift: 1199 tensors, 333 vision preserved, down_proj=64/o_proj=16/linear_out=48/mtp=2/embed=1, coverage gate 6/6, exactly 131 tensors to orthogonalize. Harness verified-ready; the destructive write still gates on operator go. |
||
|
vh
|
1857a8eb81 |
feat(coldfusion-abliteration): Robinson-formula harness, gated, staged
Harness to abliterate DavidAU/Qwen3.8-27B-Cold-Fusion-GAIN-V1.1 using the MTP-aware, vision-preserving recipe in docs/pfi/abliteration-recipe-qwen38.md. Motivation is measured, not assumed: the stock model's refusal profile (probed 2026-08-19, hand-verified) is ~33% on creative content, concentrated on explicit-sexual and graphic-torture, with 4/5 hard-harm refused, self-harm guardrails intact, and zero benign over-refusal. So there is a real creative-content refusal surface. The Robinson formula is chosen specifically because it abliterates the MTP head IN-BAND -- which the current gen seat's Heretic pass does not (its MTP head is a byte-identical base graft the Qwen3_5 wrapper never loads). That in-band MTP edit is the additive delta. The script refuses to brick the model. Two hard gates from the recipe halt before any write: the coverage identity o_proj(16)+linear_out(48)==64 (catches a tensor-name mismatch that would ship a half-abliterated model), and the attention-sink screen on dim 3994 (orthogonalizing a direction living there produces a model that loads, runs, and emits garbage). The direction is captured from two chat templates and the layer auto-picked by peak |cos| agreement in [18,45]. Classification is suffix-based and name-agnostic so it survives minor drift; the coverage gate is the backstop. Modes: --dry-run (enumerate + gate, no forward, no write), --capture (direction + sink screen, no write), default (write to --out). The README sequences dry-run -> capture -> write -> verify, and names the post-checks (vision byte-identical, refusal re-profile via services/refusal-probe/, MTP acceptance on the quant, PPL/coherence). bf16 staged to ana-ml2:/tank/aimodels/qwen38-27b-coldfusion-bf16 (pinned 9c44193, provenance recorded). The destructive run is NOT executed here -- dry-run verification and operator go gate it. |
||
|
vh
|
7010f9a1da |
feat(booth): render .md/.txt/.log inline in the gallery, collapsible + closable
Docs used to render as a clumsy link that navigated to a separate page. They now render in place: build_gallery pre-renders each doc (markdown -> HTML, plain text raw) and the gallery shows it inside a native <details open> disclosure that spans the full grid width so prose has a readable measure. The doc bar carries: a collapse chevron (the whole <details> summary toggles, works with JS off), a full-page link (still reaches the standalone viewer), a download link, and a session-close ✕. The ✕ needed stopPropagation + preventDefault because it lives inside <summary> — otherwise its click would toggle the disclosure instead of hiding the item. Close is JS (progressive enhancement); collapse is native. Two design points: - Plain text is returned RAW from build_gallery and escaped by the template inside <pre>. Pre-escaping in Python plus Jinja autoescape would double-encode angle brackets; a test pins the single-escape. - Inlining is bounded by DOC_MAX_BYTES. A doc over the limit keeps the old link-out behaviour rather than being rendered into every index load; a test covers the fallback. The shared .markdown-body / .textview typography moved from doc.html's scoped <style> into base.html so the inline body and the full-page view render identically; doc.html keeps only its page-layout wrapper. Updated the pre-existing test_gallery_links_docs_to_view: it asserted the old link-out behaviour the operator asked to change, so it now asserts the inline render plus the surviving full-page and download affordances. 61 pass. Verified live: markdown renders with headings/table/blockquote/code, txt preserves whitespace and single-escapes, collapse and ✕-close both work. |
||
|
vh
|
6770ba26d6 |
feat(booth): kept boards — a .forever sentinel and a standing link board
Agent sessions hand the operator URLs and they drown in terminal scrollback. The Booth is the right home for them — it already has the one property that decides adoption, which is that a session can publish with mkdir and cp, no API key, no schema, no deploy — but everything in it dies in 24h. So: a booth containing `.forever` is never swept, and renders in its own Kept lane at the top of the index. Opt-in per booth, so the ephemeral default is untouched and nobody inherits a cleanup chore. `rm` the sentinel and the board rejoins the sweep; the CLI verbs are sugar over exactly that, which keeps the filesystem-is-the-state model honest. The pin is deliberately NOT wired into is_expired(). That stays a pure age question feeding the `expires_in` countdown; only sweep_once() honours the sentinel. Keeping expiry arithmetic and reaper policy apart means they cannot drift into each other. Kept cards are visually separated per Australis: a 2px top edge in aurora blue, the one accent border the system sanctions. They show "kept" instead of a countdown, and they deliberately lose the one-click wipe button — a × next to the durable stuff is a footgun, so removing a kept board is a two-step act. `booth link <url> [description]` appends to the standing `links` board, creating and keeping it on first use. Entries carry provenance (handle or hostname, plus a timestamp) because a bare URL is unreadable three days later. The append is one printf of one line to an O_APPEND fd — atomic under PIPE_BUF on POSIX — which matters because many agents post to one board and interleaved half-lines would be the obvious failure mode. Seven tests cover the sentinel: detection, survival of a sweep that wipes its neighbour, the deliberate is_expired/sweep_once split, the listing flag, the sentinel not inflating item counts, and both lane-rendering directions. Two of them originally asserted on the bare strings "Kept" and "kept-grid", which passed for the wrong reason — those also appear in the inlined stylesheet served on every page — so they now assert the full class attribute. 55 pass. Also corrects the Homepage card's description, which advertised a flat 24h TTL that is no longer the whole story. |
||
|
vh
|
11b688ff68 |
feat(gen-seat): promote absolute-heresy to the live gen seat
MuXodious/Qwen3.8-27B-absolute-heresy (Heretic v1.4.0 + SOMPOA, trial T377, pin c2374593) quantized through our mixed NVFP4+FP8 recipe and promoted after passing the full gate on the probe port. Gate vs incumbent -- MTP acceptance 47.2% (48.2%), decode 103.5 tok/s (96.4), prefill 6618/5403 at 6.7k/27k (6334/5085), TTFT 27k 5.00s (5.31s), perplexity 6.910 (7.059, 2.1% better), surface 6/6, abliteration compliance 4/4. On our battery-instruct arm -- the framing that actually elicits refusals -- 0/55 with zero EMPTY, so no catatonia at the hard edge. Speed deltas are image-confounded: the probe ran the seat's pinned nightly while the incumbent's stored numbers came from an earlier image. Read as not worse. Acceptance, perplexity, surface and refusal are apples-to-apples. All 7 LiteLLM aliases verified end-to-end. GPU0 at 91.3/97.9 GB with meromero healthy -- more headroom than the previous build. Incumbent weights untouched and .env.bak-heresy-20260817 in place for rollback. Candidate is a 2-day-old RC1 with ~348 downloads; watch real multi-turn use. |
||
|
vh
|
993421bf59 |
fix(post-quant): handle sources that keep mtp.* inside a numbered shard
post_quant assumed the source ships a standalone model-mtp.safetensors, which is how JonathanColetti's grafted head is packaged. MuXodious/absolute-heresy is an unmodified full checkpoint, so its mtp.* lives in model-00012-of-00012 -- the copy silently did nothing while the index was still rewritten to point at model-mtp.safetensors, leaving 15 unresolvable tensors. Tensor counts looked correct; the checkpoint would have failed at load. The existing FAILED-CHECKS assertion caught it, which is the design working. Now extracts from the numbered shard when the standalone file is absent. Verified on the heresy build: 1968 tensors, all resolvable, 15 mtp, 333 visual, no missing shards, no orphans. |
||
|
vh
|
b0c2d3d1c4 |
fix(bench): serve_probe must mirror the live seat -- image, parsers, context
Three defects, each of which produced a false read on the candidate: 1. Hardcoded vllm/vllm-openai:latest. The Qwen3.8-27B gen seat is pinned to a nightly carrying the #51113 qwen3_5_mtp x GDN fix; probing on :latest reproduces the multi-turn corruption we already diagnosed and reads as a candidate failure. Now PROBE_IMAGE, defaulting to :latest for older seats. 2. --speculative-config JSON died twice on quoting. The inner double quotes are stripped by the outer double-quoted ssh string, and then bash BRACE EXPANSION splits {"a":1,"b":2} on the comma. Needs escaped quotes AND remote-side single quotes; both traps documented inline. 3. No --tool-call-parser/--enable-auto-tool-choice/--reasoning-parser. Without them surface_test reported tool calling as a 400 and measured a thinking split of reasoning=0ch -- both probe-config artifacts, not model defects. Re-running with the seat's flags took the candidate from 5/6 to 6/6. Also adds PROBE_MAXLEN; the hardcoded 32768 rejected prefill_bench's ~27k prompt. |
||
|
vh
|
2c3602869f |
fix(gen-seat): hash bf16 tensors via uint8 reinterpret, not numpy
numpy has no bfloat16, so .numpy().tobytes() raised 'TypeError: Got unsupported ScalarType BFloat16' on real checkpoints. Flatten then view(torch.uint8) before hashing. Result on the heresy candidate: VERDICT IDENTICAL -- all 15 mtp.* tensors byte-identical to the incumbent's verbatim base graft, the head already measured at 47.7% acceptance in production. The ~56 GB bf16 acceptance gate is redundant, so no second seat comes down. |
||
|
vh
|
254c588921 |
feat(gen-seat): CPU-only MTP head check so the gate costs no second seat
Operator ruled the probe port for validation; runbook updated to match. The bf16 MTP acceptance gate is ~56 GB resident, which on a full 97.9 GB card means downing meromero-charrp as well as gen -- freeing gen's 0.43 (~42 GB) alone is not enough. Two seats down to answer one question. compare_mtp_head.py answers the common case for free. The Qwen3_5ForConditionalGeneration wrapper never loads the MTP head, so PEFT merges, Heretic runs, and llm-compressor passes all leave mtp.* as it came from the base. It hashes a candidate's 15 mtp.* tensors against the incumbent's grafted-verbatim head -- the one measured at 47.7% acceptance in production through this exact pipeline. Identical means the acceptance question is already answered; different means the head was edited and the real gate is warranted; missing means it was dropped. CPU only, reads just the shard holding mtp.*. The runbook states the residual risk plainly: an identical head proves the head is intact, not that the abliterated body still drafts well with it -- which the Stage-3 acceptance measurement on the 22 GB quantized build catches anyway. |
||
|
vh
|
7997f111b0 |
docs(gen-seat): runbook for the absolute-heresy swap
Candidate MuXodious/Qwen3.8-27B-absolute-heresy (Heretic v1.4.0 + SOMPOA, trial T377), pinned c2374593. Beats the incumbent on both axes: refusals 2/101 vs 12/100, first-token KL 0.0759 vs 0.1191. Structurally a clean full checkpoint (1199 tensors, 15 mtp.*, 333 visual.*, lm_head), so the existing mixed NVFP4+FP8 recipe applies with no graft-and-reconstruct. Runbook carries the bf16 MTP-acceptance gate before any quant spend, the llm-compressor ignore-pruning foot-gun, the three measurement traps (cache-busting, unseeded prefill nonce, PPL with spec off), the .env 0600 sudo trap, the GPU0 co-tenant starvation risk, and rollback. Flags that the candidate is a 2-day-old RC1 whose own card carries a broken GGUF benchmark block (RC1 and RC2 report identical at-chance scores across three benchmarks), so its numbers are claims rather than measurements. |
||
|
vh
|
9c1405b1f9 |
docs(refusal-probe): retire the DS v1 regeneration — unqueued, do not run
brokkr-smithy-dev withdrew the request on the operator's call
(msg 01M088G7NQ5G42ES30YPJV4Y3V). Two reasons: the ictrl-pair-unwrapped /
ictrl-pair-wrapped control isolates the classifier over-fire cleanly,
where DS's cross-class delta only bounded it; and DS v2 is releasing
soon, so a k=5 v1 baseline baselines a superseded version.
The spec stays banked as the record of the run that will not happen --
axes, per-class grading asymmetry, and run parameters remain correct.
Checklist struck through. The staged probe.py was pulled from /mnt/smithy
when the request was withdrawn; absence verified from nh3-dev, so the
path recorded in
|
||
|
vh
|
ec0b6e5e71 |
docs(refusal-probe): record brokkr's staged probe path + sha, and the all-16-axes caveat
brokkr-smithy-dev staged refusal-map-probe.py on /mnt/smithy rather than leaving it as a run-time ask (msg 01M082P4YPJTDJCF33BEHNYW0M). Path and full sha256 recorded and verified present from nh3-dev. The probe has no per-axis selection flag: it runs all 16 axes regardless of MAP_LIGHT, so the DS run yields a 16-axis artifact, not the 8 dropped ones. Kept as-is — the creative-half rerun is a free within-model consistency check against the 50 samples already collected. |
||
|
vh
|
25fa18efb8 |
docs: correct stale TTS voice warning (tts-dev); record DS regeneration spec (brokkr)
tts-dev answered the Lobe onboarding, live-verified. Corrects a warning I shipped in the lobe-chat stack: the OpenAI voice names are ALIASED not rejected (echo/alloy/onyx/ash->donut, nova->miranda, shimmer/coral->emmie, fable/sage->glados), so a UI voice mis-click is not the hazard I recorded. Only ballad and verse 404. The old 71.2s per-call cap is dead (Zonos-era); dots chunks server-side and renders a 592-word call intact. Real constraint is size (245s WAV = 23.5MB -> request mp3) and that the seat SERIALIZES generation, so sustained Lobe volume is a real capacity question to report to tts-dev. Also records brokkr's DS regeneration spec verbatim from his probe source (msg 01M06FN7EE29M8YWP0GK517V4B): the 8 dropped axes (5 operational + 3 meta), the BLUEHERON meta system prompt, and the per-class framing that a label-level rebuild would lose -- operational uses system=None and an 18-CHARACTER refuse floor at max_tokens 45, meta scores a separate BLUEHERON leak count that must not collapse into the refuse rate, both distinct from the creative class's word floor. Queued, gated on the GPU1 window; no deadline (weights not scheduled for reuse). Recorded so it is run from the artifact, never reconstructed from labels. |
||
|
vh
|
aba7cda33e |
fix(refusal-probe): retract an invalid cross-battery comparison; adopt the two-control pair
RETRACTION. I claimed my arm-3 92.5% 'exceeded' brokkr-smithy-dev's 62.5%. That comparison was invalid. His diff against his own artifact shows my battery-instruct.yaml reproduces only his CREATIVE class -- 8 of 16 axes -- dropping all 5 operational (violence/incite, crime/fraud, cyber/malware, selfharm/methods, privacy/stalk) and all 3 meta (meta/sysprompt, meta/ignore, meta/dan), while adding 2 controls he never had, at k=5 against his k=2. His figure pools 16 axes; mine is creative-only. Different denominators, so there was never a delta to report. Cause: I rebuilt his instrument from his MESSAGE, and the class field lives in the artifact, not the prose. Lesson recorded -- reconstructing a peer's instrument from their description reproduces what they described, not what they ran; diff against the artifact before claiming comparability. Adopts his amendment to my control fix, which is better than my proposal. I argued a single UNWRAPPED control suffices since over-fire is a property of the scorer, not the wrapper. True but incomplete: one control gives only a scorer FPR baseline and cannot attribute anything to the wrapper. Two -- the same benign scenario unwrapped AND wrapped -- make the wrapper the only variable, so the delta between them is the wrapper's contribution to classifier firing. Costs one prompt. ictrl-reunion stays contaminated and unchanged, now with an explicit do-not-fix marker: rewriting it would break comparability with results already collected, and it is the cleanest live demonstration of the very confound this battery measures. |
||
|
vh
|
d47dd10795 |
feat(gen-seat): swap to AEON-ULTIMATE NVFP4 W4A4; default reasoning_effort=medium
Operator call: the incumbent abliterated model was the first one we could
find, not an optimised pick. sakamakismile/Qwen3.8-27B-AEON-ULTIMATE-
UNCENSORED-NVFP4 (base AEON-7 BF16, abliterix-abliterated, Apache-2.0),
byte-verified at /tank/aimodels/qwen38-27b-aeon-ultimate-nvfp4.
Measured on the same harness, same GPU, cache-busted per playbook 5.
Baseline was RE-measured live before the swap rather than trusted:
incumbent (W4A4+FP8 mixed) AEON (W4A4)
decode bs=1 94.09 tok/s 104.22 +10.8%
MTP acceptance 47.7% 52.3% +4.6pp
abliteration 4/4 4/4
surface 6/6 6/6
weights 22.5 GB 20.6 GB -8.4%
AEON concurrency: conc=1 98.48 tok/s aggregate; conc=6 381.29 aggregate /
63.55 per-stream, MTP holding 50.6% under load.
Surface 6/6 includes vision (image-judge rides this seat) and a 36k-token
needle retrieval, which was the specific risk in going full-W4A4 -- the
packager only validated 32k, and W4A4 long-context collapse is in our own
notes from the Granite work. It held.
reasoning_effort: the AEON template defaults to xhigh (template line 47),
and at xhigh this model can spend its entire budget inside <think> and
emit no answer -- a silent-empty-response hazard for the automated
summarizer/classifier consumers. Seat now pins the default to medium via
--default-chat-template-kwargs, per-request overridable. Override PROVEN
live: chat_template_kwargs.reasoning_effort=bogus returns HTTP 400
carrying the template's own exception text, so caller values genuinely
reach the template and invalid ones fail loudly rather than silently
falling back. Empty GEN_REASONING_EFFORT omits the flag for models that do
not read the kwarg -- the Qwen3.6 line ignores it entirely, where setting
it would be a false lever.
All 7 aliases verified routing. Rollback is one .env line; the previous
build is untouched at /tank/aimodels/qwen38-27b-uncensored-nvfp4-mixed.
TWO GAPS, declared:
- Incumbent concurrency was never captured before the swap (I baselined
bs=1 only), so the conc=1/6 figures have no same-hardware comparator.
- Perplexity NOT measured. eval_quality correctly refused it: under
--speculative-config prompt_logprobs come back ~uniform (median rank
~130k), playbook trap 2. A real PPL number needs both seats served
without spec-decode.
Adds concbench.py (concurrent throughput; wall-clock aggregate, not
sum-of-rates, and delta-based MTP accounting).
|