finding
active
finding:replication-of-anil-et-al-many-shot-sudden-learning-curves-across-five-personasReplication of Anil et al. many-shot sudden learning curves across five personas
Fig. 2 baseline.
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
- Justifies discount τ(N)=N^{-α}.
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.
- Regime with dozens to hundreds of examples where sudden learning curves are seen.
- The paper's central mechanistic explanation of why narrow fine-tuning causes broad misalignment
- Demonstrated transformers on mathematical understanding and logic; cited to motivate transformer versatility.
- Demonstrates that persona space is low-dimensional
- Characterizes what is on the far end of the Assistant Axis away from the Assistant
- Comprehensive model comparison showing tuning benefit for persona fidelity
- Systematic identification of multiple coexisting persona vectors in two open-source models
- Latents are specialized to different modes of misalignment, explaining diverse misalignment profiles