AI Paper Atlas

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

OSVBench: Benchmarking LLMs on Specification Generation Tasks for Operating System Verification

Shangyu Li, Juyong Jiang, Tiancheng Zhao, Jiasi Shen

AAAI 2026 · Large language models › Agents, code and math · approach: Execution-based code verification · 17 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 few-shot syscall examples (0-shot to 1/3/5-shot) (ablation)five LLMs on OSVBench · OSVBenchPass@1: 0% at 0-shot, large gains with more shotsImproved
increase number of demonstrations from 1 to 5 (ablation)DeepSeek-Chat, o3-mini and others · OSVBenchPass@1Improved
inject bugs into the syscall code in the prompt (ablation)12 LLMs · OSVBenchPass@1: 53 of 60 (88.3%) bug experiments lower than bug-freeHurt
increase number of injected bugs per task (ablation)12 LLMs · OSVBenchPass@1Hurt
use reasoning model (o1) instead of GPT-4oOpenAI models · OSVBenchPass@1 total: 38.78 -> 23.67Hurt
use reasoning model DeepSeek-R1 instead of DeepSeek-ChatDeepSeek · OSVBenchPass@1 total: 46.53 -> 40.82Hurt
use larger model (70B/72B vs 7B/8B)Llama-3.1 and Qwen2.5 · OSVBenchPass@1 total: 10.61 -> 22.45 (Llama); 4.90 -> 32.24 (Qwen)Improved
increase k in Pass@k from 1 to 3 to 5 (ablation)GPT-4o, DeepSeek-Chat, Llama-3.1-70B-Instruct · OSVBenchPass@kImproved
add self-repair with verifier error messages (ablation)GPT-4o and DeepSeek-Chat · OSVBenchRepair success rate: 14.67 (round 1), 1.56 (round 2) for GPT-4oImproved
add a second self-repair round (ablation)GPT-4o and DeepSeek-Chat · OSVBenchRepair success rate: 14.67 -> 1.56 (GPT-4o); 11.45 -> 2.59 (DeepSeek-Chat)No clear effect

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