Unverified50% confidenceBenchmarkExact time
The Alpha Codium method achieved significantly higher accuracy on the code competition benchmarks HumanEval and CodeContests compared to direct prompting (zero-shot).
1
Sources
50%
Confidence
Long-term
Relevance
10/7/2026
First Seen
Sources
Related Entities
Related Claims
Unverified一项代码生成研究发现，分阶段提示的准确率比一次性提示高得多68% similarUnverifiedCode correctness can obtain a deterministic 0/1 signal through unit testing, and this high-quality, low-cost reward signal makes reinforcement learning far more efficient in code scenarios than in natural language scenarios.67% similarUnverifiedExecution-based Reward signals derived from unit test execution results are significantly effective in improving code correctness rates.67% similarUnverifiedThe core idea of Alpha Codium is to replace single-prompt code generation with a test-driven iterative process, in which the model first understands the problem specification, generates public and private test cases, and then iteratively refines the code until all tests pass.66% similarUnverified当前主流基准测试体系如HumanEval、MBPP等代码评估集往往关注平均性能，对分布式的场景特定退化现象敏感度不足65% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/986062API
curl https://kongchang.com/api/v1/knowledge/claims/986062MCP
get_claim(id=986062)