CATCH Benchmark

What is CATCH?

During reinforcement learning with verifiable rewards (RLVR), large language models (LLMs) can exploit loopholes in their environments to obtain high rewards…

Released

2026-09-30

Evaluates

Cybersecurity, Language & Knowledge, cs.CL

Openness

unknown

Importer

Claire Radar

Review status

rule-auto-admitted

Reported scores

0

Source provenance

• Original evidence https://arxiv.org/abs/2609.39533

CATCH paper and code

• Paper https://arxiv.org/abs/2609.39533
• Code https://github.com/THUAIS-Lab/CATCH

Which sources cite CATCH?

• CATCH

Related Cybersecurity, Language & Knowledge, cs.CL benchmarks

• 2d-fet-bench
• 5-Dialects-BN
• A Large-Scale Dataset and Benchmark
• AARRI-Bench
• ABC-Bench
• ActiveSciBench
• AdaMem-Bench
• AdaPlanBench