Unverified50% confidenceOpinionExact time
The difference between REPA and RECON is that REPA is a soft constraint at the loss level (making backbone features fit encoder features), while RECON is hard information at the input level (feeding encoder features directly into the backbone as input).
1
Sources
50%
Confidence
Long-term
Relevance
10/15/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedREPA's alignment loss acts only on the early layers, while the later layers are supervised only by the velocity prediction loss and lack a mechanism for transmitting semantic information, which is REPA's structural flaw.72% similarUnverified"Rebasing" means that when a report is rebuilt due to staleness, each assumption's baseline value is updated to the current live value, and the next comparison starts from there. This avoids error accumulation, at the cost of losing a global view of long-term drift.60% similarUnverifiedCommon solutions for vanishing gradients include replacing sigmoid with ReLU and its variants, batch normalization, and residual connections (the core idea behind ResNet).60% similarUnverifiedRerank is the fine-ranking stage after multi-way retrieval in RAG, used to more finely score the relevance of candidate documents to the question.59% similarUnverifiedRebuttal can improve scores but has limited effect; its core value lies in clarifying reviewers' misunderstandings, supplementing experimental data, and resolving reproducibility disagreements.57% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1037873API
curl https://kongchang.com/api/v1/knowledge/claims/1037873MCP
get_claim(id=1037873)