docs: Jev candidate bench vs SemIf (fv-ml1 GPU 3) — Intern-Decision-4B is the replacement if SemIf is displaced

Operator ask relayed by brokkr-smithy-dev. Positive control (SemIf 187/231, hard 0.613) reproduced exactly; negative control and a 4-restart noise floor (0 flips) measured. On our replaced-baseline sets no candidate beats SemIf-with- rotations beyond the ~4-pt floor; Intern-Decision-4B native matches it at one ordering, is better on Wyrd, fits 9.7/10.3 GB and is 1.5-2.3x faster. JevBench rank does not transfer. Raw per-item data kept out of git.
2026-09-30 05:00:45 -07:00
parent 9a6ac59da7
commit 475d6d6bcb
27 changed files with 19068 additions and 0 deletions
@@ -47,3 +47,6 @@ stacks/lobe-chat/.env
 # and every commit here is attributed to Vuong Hoang anyway, so git could
 # not carry the attribution this file exists to provide.
 .ops-log/
+
+# Jev bench raw per-item responses (83 MB): kept on disk (restic via /home) and in the Booth, not in git
+services/semif-serve/bench-jev-2026-09-30/raw/