SEVRA-BENCH

What is SEVRA-BENCH?

SEVRA-BENCH is a benchmark for measuring how often LLM-based code review agents approve adversarial pull requests with social-engineering framings, built from…

Released

2026-06-11

Evaluates

Agents, Cybersecurity, Language & Knowledge, Robotics & Embodied AI

Openness

unknown

Importer

Claire Radar

Review status

unreviewed

Reported scores

0

Source provenance

• Original evidence https://arxiv.org/abs/2606.13757

SEVRA-BENCH paper

• Paper https://arxiv.org/abs/2606.13757

Which sources cite SEVRA-BENCH?

• SEVRA-BENCH

Related Agents, Cybersecurity, Language & Knowledge, Robotics & Embodied AI benchmarks

• $\tau^\tau$-Bench: An Environment for End-To-End, Realistic Agent Construction
• 360CityArena
• 4DSynth-Nav
• 5-Dialects-BN
• A Large-Scale Dataset and Benchmark
• A3S-Bench
• AARRI-Bench
• ABC-Bench