claim
active
claim:simple-interpretability-methods-scale-to-2-8t-parameter-models-and-long-context-agentic-rolloutsSimple interpretability methods scale to 2.8T-parameter models and long-context agentic rollouts
Scaling claim.
Source paper
extracted_from(2026) · Leon Bergen · Usha Bhalla · Andrew Lee · Barak Widawsky +14
Neighborhood — ranked by edge-count
Papers (1)
paper
Findings (1)
finding
- Probe predictive of downstream hacks.
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.
- Abstract.
- Broadly, turn-averaged SAEs make interpretability techniques practical at long context lengths.quote0.772Abstract thesis.
- Positioning claim of the release.
- A better basis (e.g., factoring tokens into languages or parts of speech) may yield a sparse interpretable modelhypothesis0.770Same model can be dense in one basis and sparse in another.
- Authors' outlook.
- Motivation for the release.
- Motivates shift from studying model activations ('thoughts') to understanding parameters ('the computations themselves').
- Claude 3 Opus ratings aligned with human judgment of feature descriptions.