Unverified50% confidenceBenchmarkExact time
The authors evaluate on two code generation benchmarks, HumanEval and MBPP, and the results show that both the practical decoding algorithms based on these decision rules and CTD outperform standard temperature sampling and adaptive decoding baselines.
1
Sources
50%
Confidence
Long-term
Relevance
10/14/2026
First Seen
Sources
Related Entities
Related Claims
Unverified传统代码评测基准（如HumanEval、MBPP）主要测试模型在孤立函数级别的生成准确率73% similarVerifiedCode large language models are evaluated using executable benchmarks such as HumanEval and MBPP, meaning the generated code is directly executed and the output is checked for correctness.73% similarVerifiedModel evaluations in the code generation field, such as HumanEval and SWE-bench, place special emphasis on functional correctness rather than fluency.71% similarUnverifiedHumanEval, MBPP, and SWE-bench are standard benchmark datasets specifically targeting code generation.71% similarUnverifiedCodex模型的评估通常使用HumanEval和MBPP等编码基准测试68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1029444API
curl https://kongchang.com/api/v1/knowledge/claims/1029444MCP
get_claim(id=1029444)