Unverified50% confidenceFactExact time
Scaling data according to the research findings, the trained agents outperform models trained on data from prior work on all four evaluation benchmarks.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/28/2026
First Seen
Valid until: 1/26/2027
Sources
Related Entities
Related Claims
Unverified小模型通过知识蒸馏通常比直接在标注数据上训练精度高出2-5个百分点78% similarUnverifiedMultiple studies in 2024 found that the strong performance of certain models on standard benchmarks resulted from benchmark test questions being mixed into the training data (data contamination).78% similarUnverified智能体的偏好受训练数据、提示词、版本迭代等多重因素影响，今天有效的优化手段可能在下一次模型更新后失效74% similarUnverifiedModels trained at large scale may acquire meta-cognitive information from training data about 'being an AI' and 'potentially being under evaluation'.74% similarUnverifiedLarge language model-driven agents tend to select the tools that appear most frequently in training data and are most convenient for completing tasks, rather than proactively weighing whether those tools meet data security requirements.74% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1106012API
curl https://kongchang.com/api/v1/knowledge/claims/1106012MCP
get_claim(id=1106012)