Unverified50% confidenceFactExact time
Cross-modal alignment still lags substantially behind the alignment level between models of the same modality.
1
Sources
50%
Confidence
Long-term
Relevance
11/10/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedAlthough cross-modal alignment is advancing, it still has a clear gap with same-modality alignment, and explicitly aligned multimodal models are not the end point.77% similarUnverified偏差-方差权衡原理表明单一模型难以同时最小化偏差和方差,通过聚合多个具有差异性的模型可以突破这一限制73% similarUnverifiedProgress in alignment research is widely considered to lag far behind the pace of improvement in model capabilities themselves.73% similarUnverifiedUnder the assumption that the state transition Jacobian decays exponentially, the error between the truncated gradient and the full-context gradient decays exponentially as the distance from the context boundary increases.72% similarVerifiedSmaller models often show exponential rather than linear degradation in performance on long reasoning chain tasks.71% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1154510API
curl https://kongchang.com/api/v1/knowledge/claims/1154510MCP
get_claim(id=1154510)