AI Paper Atlas

Research map / Large language models / Agents, code and math

Can LLMs Reason Like Automated Theorem Provers for Rust Verification? VCoT-Bench: Evaluating via Verification Chain of Thought

Zichen Xie, Wenxi Wang

ICML 2026 · Large language models › Agents, code and math · approach: Formal verification and proof data · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add thinking mode (ablation)Qwen 3 8B · VCoT-Bench-Type (Invariant removal)overall accuracy (%): 26.87 -> 10.03Hurt
add thinking mode (ablation)Qwen 3 8B · VCoT-Bench-Type (Assertion removal)overall accuracy (%): 13.70 -> 11.47Hurt
use Gemini 3.1 Pro instead of Gemini 3 FlashGemini · VCoT-Bench-Loc (Middle)overall accuracy (%): 47.69 -> 36.15Hurt
move missing blocks from Front to MiddleClaude Sonnet 4.5 · VCoT-Bench-Locoverall accuracy (%): 69.04 -> 46.35Hurt
move missing blocks from Front to MiddleGemini 3.1 Pro · VCoT-Bench-Locoverall accuracy (%): 62.69 -> 36.15Hurt
increase removed proof blocks from 10% to 100% (ablation)Claude Sonnet 4.5 · VCoT-Bench-Ratiooverall accuracy (%): 71.58 -> 17.22Hurt
increase removed proof blocks to 100% (ablation)Qwen 3 think · VCoT-Bench-Ratiooverall accuracy (%): 0.66Hurt
use Best-of-10 sampling instead of greedy decodingGPT-5.2 · VCoT-Bench-Type (Assertion removal)overall accuracy (%): 34.25 -> 45.03 (+10.78)Improved
use Best-of-10 sampling instead of greedy decodingDeepSeek V3.2 · VCoT-Bench-Type (Invariant removal)overall accuracy (%): 49.49 -> 50.00 (+0.51)Improved
use GPT-5.2 instead of Claude Sonnet 4.5 for Z3-to-Verus transformation (ablation)VCoT-Lift · 20 Verus-Bench programsavg lifted proof lines: 7 -> 45Improved
use Gemini 3.1 Pro instead of GPT-5.2 for proof repair (ablation)VCoT-Lift · 20 Verus-Bench programsavg repair rounds: 13 -> 4Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy