SWE-DeceptionBench: Benchmarking Deceptive Behaviors of LLMs Across the Software Development Lifecycle

arXiv:2610.13626 · cs.SE, cs.AI · Submitted 2026-10-09 · Read on arXiv

cs.SE, cs.AI

Submitted: 2026-10-09

Updated: 2026-10-09

Code: https://github.com/THU-Agent/SWEDeceptionBench

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Terminology

• Benchmarking — 1× in this paper · explained in 4 episodes

Sources

• Evaluating Large Language Models Trained on Code
• The Llama 3 Herd of Models
• Frontier Models are Capable of In-context Scheming
• The MASK Benchmark: Disentangling Honesty From Accuracy in AI Systems
• From Sycophancy to Deception: A Unified Taxonomy for LLM Spontaneous Misalignment
• Qwen3Guard Technical Report

Related papers

• Falsification-Based Verification of LLM-Generated Optimization Models: Sound Test Batteries and Their Detection Limits
• GitSkills: A Dataset of Agent Skills on GitHub
• SABER: Benchmarking Operational Safety of LLM Coding Agents in Stateful Project Workspaces
• PackMonitor: Enabling Zero Package Hallucinations Through Decoding-Time Monitoring
• IntentCoding: Amplifying User Intent in Code Generation
• Incentives and Outcomes in Bug Bounties