Unverified50% confidenceFactExact time
The post-training stage carries the risk of catastrophic forgetting, and post-training of web-native models usually adopts a mixed-data strategy, training on a blend of general instruction data and web interaction data.
1
Sources
50%
Confidence
Long-term
Relevance
10/12/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedWhen the distribution of online data deviates from the training data, model performance degrades (data drift and model decay), requiring retraining to be triggered.77% similarUnverifiedMainstream benchmark problems have been widely discussed on the internet, and models' training data may contain memorization of these problems, causing benchmark scores to overestimate the models' capabilities on truly unseen tasks73% similarUnverifiedThe core assumption of the research is that shared training data and similar architectures cause more capable LLMs to exhibit higher behavioral consistency with one another, thereby generating systemic risk73% similarUnverifiedMethods 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.73% similarVerifiedResearch by Google DeepMind and other institutions in 2023 confirmed that a model's memorization behavior is closely related to the number of repetitions of training data.73% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1006641API
curl https://kongchang.com/api/v1/knowledge/claims/1006641MCP
get_claim(id=1006641)