Reasoning models internally represent step importance in activations prior to generation, detectable via probes and independent of token-level features.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Reasoning Models Know What's Important, and Encode It in Their Activations
Reasoning models internally represent step importance in activations prior to generation, detectable via probes and independent of token-level features.