Experiment ·
CUT7
Does adding FACTS r6 to the frozen cut 6 build change any v0.2 held-out or V-para answer, and if so are changed answers better with zero regressions and zero wrong fires (control = cut 6 claim-run traces/verdicts)?
From the record
- What was tested
- CUT7
- Result
- 2026-10-06 rule: only routing/safety fixes with v0.1 answers unchanged; 2026-10-07: cut 7 = cut 6 + FACTS r6 changed 8 v0.1 answers (0 fires on the held-out sets)
- Verdict
- adopted: shipped or recommended by its frozen rule
- Detail
- lever-adopted (claim build cut 7, adopted on gates alone under the zero-fire clause, CUT7-criterion.md:11)
- Set
- v0.2 (bench-v02); V-para (V-para); v0.1 public + safety track; dev chat 192; V-exact 11
- Judge
- none (deterministic gate)
- n
- v0.2 110, V-para 58, v0.1 public + safety 72
- Seeds
- s42 queries (s43 verified identical); no LLM traces
- Pre-registered
- yes, sha256-frozen
- Reopen if
- recall axis only: see FACTS-RECALL.md and LCRELAX-PROBE.md
Linked facts · 2
- The cut 6 freeze was relaxed once: cut 7 added the Wikidata fact table, which changed 8 public dev answers.The cut 6 freeze was relaxed once: cut 7 added the Wikidata fact table, which changed 8 public dev answers.2026-10-06 rule: only routing/safety fixes with v0.1 answers unchanged; 2026-10-07: cut 7 = cut 6 + FACTS r6 changed 8 v0.1 answers (0 fires on the held-out sets)Setv0.1Statuspublic devJudgenot recordedn8 of 57 v0.1 answersSeedsnot recordedDate2026-10-07Sourcenot in the public repositories yetOpen the data file
- R9 FACTS instant tier: Wikidata table r1–r6R9 FACTS instant tier: Wikidata table r1–r6recall 70 % < 75 % barKinda record, not a measurement (set, status, judge, n, seeds, date: not recorded)Caveatwhy closed: FAIL by frozen rule; shipped by decisionSourceresearch record
reports/CLOSED.md:24Open the data file