Unverified50% confidenceFactTime unknown
The Needle-in-a-Haystack test randomly inserts an irrelevant target sentence at different positions within a long text, and by querying the model draws a two-dimensional heatmap of 'position × context length' to present retrieval success rates.
1
Sources
50%
Confidence
Long-term
Relevance
9/12/2026
First Seen
Sources
Context Degradation Benchmark: The Long-Context Weakness of Leading LLMs
githubkyleaoconnell22
Related Entities
Related Claims
UnverifiedThe Needle-in-a-Haystack test randomly inserts key information at different positions within an ultra-long document and tests whether the model can accurately locate it, making it a standard stress test for evaluating long-context capabilities.77% similarUnverifiedAn ideal benchmark should fix the model, tools, and task distribution, varying only trajectory length as a single variable.68% similarUnverified跑分排名不等于所有场景的实际表现，具体任务仍需具体测试68% similarUnverifiedThe experiments were tested across six heterogeneous benchmarks.67% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/917853API
curl https://kongchang.com/api/v1/knowledge/claims/917853MCP
get_claim(id=917853)