Segment-level supervision extracts coherent proof segments to train policy models that achieve 61-66% success on miniF2F, outperforming step-level and whole-proof methods while also improving existing provers.
The lean 4 theorem prover and programming language
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.AI 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Rethinking Supervision Granularity: Segment-Level Learning for LLM-Based Theorem Proving
Segment-level supervision extracts coherent proof segments to train policy models that achieve 61-66% success on miniF2F, outperforming step-level and whole-proof methods while also improving existing provers.