Unverified50% confidenceFactExact time
An arXiv paper (arXiv:2609.35890v1) finds that representation-level sparsity and circuit-level causal sparsity do not always move in sync, and can diverge at specific faithfulness thresholds.
1
Sources
50%
Confidence
Long-term
Relevance
9/30/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedAn arXiv paper (arXiv:2609.26986v1) points out that the sequential order in which evidence appears is a long-underestimated confounding factor when measuring the modality preferences of multimodal large models.67% similarUnverifiedAn arXiv paper (arXiv:2609.30288) proposes using low-rank bottleneck autoencoders to obtain context-mixing capabilities similar to attention while reducing computational overhead.65% similarUnverifiedAn arXiv paper (arXiv:2609.15990v1) spanning 12 open-weight models, 2 task types, and 2 architectures investigates the problem of few-shot prompting sometimes causing model performance to degrade.62% similarUnverifiedThe paper points out that the root cause of long-context degradation lies not entirely in context length, but in the number of effective distractors it contains.62% similarUnverified西尔维斯特-加莱定理依赖于实数域的特殊性质，并非纯粹的组合结论，在复平面上失效61% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/961067API
curl https://kongchang.com/api/v1/knowledge/claims/961067MCP
get_claim(id=961067)