On CoTemp QA, supervised fine-tuning with raw R1 traces gave the best model accuracy while human raters found those traces least interpretable, showing model-useful traces and human-readable traces can diverge.
Mann-whitney u test
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.CL 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
unclear 1representative citing papers
citing papers explorer
-
Do Cognitively Interpretable Reasoning Traces Improve LLM Performance?
On CoTemp QA, supervised fine-tuning with raw R1 traces gave the best model accuracy while human raters found those traces least interpretable, showing model-useful traces and human-readable traces can diverge.