Unverified50% confidenceFactExact time
DiscoBench的463个歧义实例覆盖四种类型：实体歧义、版本歧义、标准歧义、事实错误歧义
1
Sources
50%
Confidence
Long-term
Relevance
7/9/2026
First Seen
Sources
Related Claims
Unverified使用少样本提示词模板时，2-5个高质量、有代表性的示例胜过大量低质量示例，示例的多样性和边界覆盖度比数量更重要56% similarUnverified文档分块需权衡片段大小与语义完整性，过小则语义不完整，过大则引入噪音56% similarUnverified测试误差可以分解为偏差的平方、方差和不可约噪声三个部分55% similarUnverifiedThis evaluation was designed by Bilibili UP host '不正经的前端' and includes four coding scenarios of increasing difficulty.54% similarUnverifiedIn chunking strategies, chunks that are too large cause semantic focus to blur, while chunks that are too small cause context fragmentation; fixed-window, sentence-level, and semantic-boundary recursive chunking each involve trade-offs in different scenarios, with no universally optimal solution.53% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/387219API
curl https://kongchang.com/api/v1/knowledge/claims/387219MCP
get_claim(id=387219)