wassnameandClaude Opus 4.7 a2e3124e06 fix airisk self_violate actor archetype + simplify to 2 conditions ...

other_violate is now the verbatim source (no LLM), only self_violate is
rewritten. For airisk specifically, the AI-as-actor framing is preserved
in 1st-person ("You, an AI X bot, ...") since a naive "You ..." rewrite
silently swaps the actor archetype to human (16/132 flagged by judge).

Pairwise consistency check (06_consistency.py) verifies same-situation
across the perspective shift. After the fix:
- clifford 97.7% same
- scifi    99.2% same
- airisk   86.3% -> 100.0% same

First eval signal on Qwen3-0.6B: airisk wrongness=+0.70, gap=+0.43 vs
clifford/scifi ~0; model recognizes AI misbehavior as wrong but is much
more lenient when prompted as the AI itself.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

2026-04-30 21:29:38 +08:00

..

01_download.py

init

2026-04-30 17:10:09 +08:00

02_rewrite.py

decent

2026-04-30 21:22:07 +08:00

03_eval.py

decent

2026-04-30 21:22:07 +08:00

04_validate.py

decent

2026-04-30 21:22:07 +08:00

05_drift.py

decent

2026-04-30 21:22:07 +08:00

05_upload_hf.py

fix airisk self_violate actor archetype + simplify to 2 conditions

2026-04-30 21:29:38 +08:00

06_consistency.py

fix airisk self_violate actor archetype + simplify to 2 conditions

2026-04-30 21:29:38 +08:00