Unverified50% confidenceFactExact time
For the curriculum-overlap benchmark Skolprov, the incremental explanatory power was ΔR²=.128 (95% CI [.038,.253]), indicating that translated MMLU scores significantly predict performance on this native-language benchmark.
1
Sources
50%
Confidence
Long-term
Relevance
10/30/2026
First Seen
Sources
Related Entities
Related Claims
Unverified代码大语言模型在HumanEval基准测试上的通过率从2022年OpenAI Codex的约47%提升到2024年顶级模型(GPT-4o、Claude 3.5 Sonnet)超过90%62% similarUnverifiedCompared with BPE, Unigram, and parity-aware BPE, LCT improves the aggregate score on four multilingual downstream benchmarks by 1.48, 1.83, and 2.00 points, respectively.60% similarUnverifiedThe average KL divergence of the UD-Q4_K_XL quantization published by Unsloth is 0.046893, with a top-1 agreement rate of 92.255% with BF16.60% similarUnverifiedThe best-performing model, AfriNLLB-12, achieved a BLEU score of 7.76 and a chrF++ score of 30.72 on the Wolof-to-Arabic direction.60% similarUnverifiedGLM-5.3的智能体编程准确率随投入词元增加而提升:低投入约24.7,高投入31.5,最大投入34.760% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1114311API
curl https://kongchang.com/api/v1/knowledge/claims/1114311MCP
get_claim(id=1114311)