Unverified50% confidenceSolutionExact time
For models trained on Danbooru, writing prompts using Danbooru tags is often more precise than using natural language descriptions.
1
Sources
50%
Confidence
Long-term
Relevance
9/27/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedEarly mainstream models such as NovelAI and Stable Diffusion were trained extensively on the Danbooru dataset and therefore have a very strong instruction response to the Danbooru tagging system.76% similarVerified大语言模型在RLHF训练阶段，因人类标注员倾向于给更详细、更礼貌的回答打高分，导致模型学会了过度冗长的表达模式61% similarUnverifiedWhen a language model is asked to show its thinking, what it outputs is a self-description reorganized from linguistic patterns related to relevant concepts in its training data, not a genuine readout of its internal states.60% similarVerifiedThe root cause of hallucination lies in the fact that the model's training objective optimizes for generating fluent and coherent text rather than generating truthful text.60% similarUnverifiedThe root cause of citation hallucination is that LLMs learn linguistic patterns by predicting the next word, and no factual database for real-time verification is embedded in the training process.59% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/954692API
curl https://kongchang.com/api/v1/knowledge/claims/954692MCP
get_claim(id=954692)