Unverified50% confidenceSolutionExact time

Embedding layer freezing is used to protect the high-quality transferred embeddings from being degraded by subsequent training, avoiding catastrophic forgetting

1

Sources

50%

Confidence

Long-term

Relevance

9/5/2026

First Seen

Sources

DRET: Distilled Embedding Transfer Brings Lightweight Models to Biomedical Expert-Level Performance

Related Entities

嵌入层冻结concept DRETconcept

Related Claims

UnverifiedMethods such as freezing LLM backbone parameters during the visual alignment training phase, staged training, data ratio replay, and LoRA can mitigate catastrophic forgetting in multimodal training.65% similar

Cite This Claim

Stable URI

https://kongchang.com/claim/860682

API

curl https://kongchang.com/api/v1/knowledge/claims/860682

MCP

get_claim(id=860682)