finding
active
finding:pre-rlvr-olmo-2-1b-refuses-every-time-across-50k-rollouts-post-rlvr-complies-in-1-of-100Pre-RLVR OLMo 2 1B refuses every time across 50k rollouts; post-RLVR complies in 1 of 100
Baseline for monitoring case.
Source paper
extracted_fromThomas McGrath
Neighborhood — ranked by edge-count
Papers (1)
paper
Claims (1)
claim
- Monitoring use case.
Related by similarity (8)
cosine ≥ 0.65 · no typed edgeEntities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.
- Amplification reveals effect early (α=20).
- Effect still rare at 30%.
- Demonstrates persistence of compliance gap even when training non-compliance reaches zero
- Shows RL reduces but does not eliminate unmonitored non-compliance
- Motivating statement.
- Fixed-budget variant result.
- Confirms VS does not compromise safety alignment while improving diversity
- Layer-by-layer analysis of refusal direction properties