Early-Signal Detector

benchmark · first seen 6 Oct 2026

CWEval

Benchmark evaluating functional correctness and security of LLM-generated code across CWE categories.

Mentions per week

0219 May11 Aug10 NovPapersCodeDiscussion

PapersCodeDiscussion

Earliest sources

Latest sources

Mentioned together with

functionality-security gapgrammar-constrained decodingin-stream self-repair of generated code