Unverified50% confidenceFactExact time
Conditional Anchored Generation Distillation (CAGD) is an approach targeting catastrophic forgetting in continual learning of language models; it uses the old model's predictive distribution as the distillation target, protecting model capabilities at the behavioral level rather than the data level.
1
Sources
50%
Confidence
Long-term
Relevance
10/7/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedCAGD decouples three roles that are conflated in data replay: the condition (selecting which behaviors to protect), teacher generation (locating where to protect), and soft targets (specifying how predictions may change).70% similarVerifiedSpeculative decoding is a lossless acceleration technique that does not alter the model's output distribution; it uses a draft model with far fewer parameters than the target model to generate candidate tokens, which the target model then verifies in parallel.66% similarUnverifiedThe arXiv paper "Learning Style, Forgetting Semantics" explains the phenomenon of SFT being more prone to forgetting than RFT through theoretical modeling.64% similarUnverifiedThe challenge of streaming model distillation lies in the fact that if the teacher model adopts a non-causal architecture, its latent representations contain future information that the student model cannot access during inference, requiring the design of alignment strategies to prevent the student from being forced to fit information it cannot reproduce.64% similarUnverifiedA jailbroken model refers to a large language model whose built-in safety filtering mechanisms have been bypassed through prompt engineering techniques or technical methods to produce restricted content.64% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/982971API
curl https://kongchang.com/api/v1/knowledge/claims/982971MCP
get_claim(id=982971)