finding
active
finding:predictive-codes-form-hierarchical-block-structures-aligned-with-sub-event-boundaries-in-stories-while-saes-cluster-tokens-by-lexical-identityPredictive codes form hierarchical block structures aligned with sub-event boundaries in stories, while SAEs cluster tokens by lexical identity.
Fig. 5.
Source paper
extracted_from(2025) · Lubana, Ekdeep Singh · Rager, Can · Hindupur, Sai Sumedh R. · Costa, Valerie +12
Neighborhood — ranked by edge-count
Papers (1)
paper
Claims (1)
claim
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.
- Fig. 8a.
- Fig. 9b.
- Fig. 7c over 50 GPT-5 stories.
- Fig. 9c, Fig. 37.
- Related framework emphasizing prediction errors; active inference extends to Markov decision processes.
- Proposition 4.1 central claim.
- Extension of mechanistic interpretability findings to the metacognitive domain
- Fig. 7d,e.