finding
active
finding:replacing-cross-validated-hyperparameters-with-fixed-values-increases-tvd-by-0-0126-for-llama-at-threshold-0-15Replacing cross-validated hyperparameters with fixed values increases TVD by 0.0126 for Llama at threshold 0.15.
Ablation.
Source paper
extracted_from(2026) · Eric Bigelow · Amir Zur · Satchel Grant · Tal Haklay +6
Neighborhood — ranked by edge-count
Papers (1)
paper
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.
- Ablation.
- Replacing kernel pooling with flat per-segment pooling raises pooled TVD by 0.0021 for Llama at S=30.finding0.795Ablation.
- Scale test.
- Illustrative finding that ESR mitigates but does not fully eliminate steering influence
- Demonstrates ESR can be deliberately enhanced through prompting in the largest model
- Multi-attempt improvement rate peaks at 83% around -1.0σ below threshold in Llama-3.3-70Bfinding0.781Shows slightly weaker steering allows more successful corrections, characterizing optimal ESR conditions
- Shows behavioral pattern of self-correction is trainable in smaller models
- Cross-judge validation of the primary ESR finding across OpenAI, Alibaba, Anthropic, and Google judge models