Files
gbrain/evals
Garry Tan 1077868e36 fix: adversarial-round hardening — close the residual gate bypass + free invariants
Claude adversarial pass (post-red-team): (1) the same-hash drift defense had
a foreign-fixtures_hash carve-out a doctored committed baseline could slip
through — removed; ANY committed-vs-main divergence must now be
receipts-backed by the actual run (foreign-corpus callers opt out via
--committed-baseline); (2) same corpus must yield identical gold_total —
free invariant now breached on any scorer/loader drift; (3) hybrid
continuity writers can no longer double-count into the write-back cell;
(4) retrieval gold on assistant turns (never replayed) is now a validation
error. Hygiene: gate script cleans its defaulted temp artifact, CI uses
runner.temp instead of a fixed /tmp path, codex preamble failures are loud,
accepted residuals (justification is human-judged, count-preserving
dilution, no auto-ratchet) documented in BRAINBENCH.md.
2026-06-12 13:03:09 -07:00
..