TokUR estimates token-level uncertainty via low-rank weight perturbations in LLMs, aggregates signals to correlate with correctness, and uses them to improve reasoning performance on math tasks.
In Appendix B, we present the full algorithmic description of our method with low-rank weight perturbation
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
TokUR: Token-Level Uncertainty Estimation for Large Language Model Reasoning
TokUR estimates token-level uncertainty via low-rank weight perturbations in LLMs, aggregates signals to correlate with correctness, and uses them to improve reasoning performance on math tasks.