finding
active
finding:overthinking-can-nearly-double-inference-cost-without-improving-accuracyOverthinking can nearly double inference cost without improving accuracy
Background finding establishing the practical cost of the overthinking phenomenon this paper explains.
Source paper
extracted_from(2026) · Jeffrey Lai · Anthony Bao · J. Quinn · William Gilpin
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.
- Future work for frequent tasks.
- Performance failure mode in recurrent networks from excessive computation iterations, relevant to looped LLM design
- Central interpretation.
- Computational expense proportional to number of generated tokens, targeted for reduction by ReflCtrl
- §2, comment on expected free energy decomposition.
- Inverse scaling interpretation.
- Interpretation of Grok 4 vs Grok 4 Fast per-koan comparison
- Smoothed behavior.