Large reasoning models frequently ignore explicitly injected ground-truth solutions mid-trajectory and continue to overthink, often ending on incorrect answers.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Large Reasoning Models are not thinking straight: on the unreliability of thinking trajectories
Large reasoning models frequently ignore explicitly injected ground-truth solutions mid-trajectory and continue to overthink, often ending on incorrect answers.