Unverified50% confidenceFactExact time
2019年Dodge等人的研究发现NLP领域大量论文缺乏足够的实现细节，导致独立研究者无法复现原始结果
1
Sources
50%
Confidence
Long-term
Relevance
7/13/2026
First Seen
Sources
Can a 2.5 Overall Score Get Accepted at EMNLP? A Deep Dive into Review Scoring
redditr/learnmachinelearning7/12/2026
Related Claims
Unverified现有科学文献实验描述缺乏机器可读标准化格式，大量实验细节依赖隐性知识，亟需实验描述本体论领域的基础建设65% similarUnverifiedMost research results in continual learning remain at the research stage, lacking real production environment deployment.61% similarUnverifiedResearchers have developed evaluation benchmarks such as BIG-Bench, MMLU, and ARC, but these benchmarks still have insufficient coverage in highly specialized engineering domains such as computer architecture design.59% similarUnverified在学术调研场景中召回率往往比准确率更重要，因为漏掉关键论文代价远大于返回噪声文献58% similarUnverified表格数据的大模型化比NLP领域晚了约5年，因为表格数据的异构性、稀疏性和缺乏统一标记化方案使技术门槛更高57% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/501573API
curl https://kongchang.com/api/v1/knowledge/claims/501573MCP
get_claim(id=501573)