finding
active
finding:llama-3-1-70b-cross-validated-r-0-98-with-icl-occurring-faster-than-smaller-modelsLlama-3.1-70B cross-validated r=0.98 with ICL occurring faster than smaller models
Scale test.
Source paper
extracted_from(2025) · Eric Bigelow · Daniel Wurgaft · YingQiao Wang · Noah D. Goodman +3
Neighborhood — ranked by edge-count
Papers (1)
paper
Claims (1)
claim
- Core thesis.
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.
- Supporting finding showing ESR is driven by both higher multi-attempt rates and comparable improvement rates
- The model used in Experiment 2 for SAE feature steering experiments via Goodfire API
- ICL dynamics on Llama-3.1-8B are sigmoidal in N^{1-α}, fit by power-law-discounted likelihoodfinding0.823Prediction 1 confirmed.
- Illustrative finding that ESR mitigates but does not fully eliminate steering influence
- Llama-3.3-70B exhibits internal consistency-checking mechanisms that operate during inferenceclaim0.818Central interpretive claim of the paper supported by causal ablation and activation evidence
- Quantitative vulnerability profile for Llama-3.1-8B showing AS dominance
- Cross-judge validation of the primary ESR finding across OpenAI, Alibaba, Anthropic, and Google judge models
- One of four LLMs selected; larger model with D=8192 embedding dimension; analyzed across proportionally aligned layers.