Unverified50% confidenceFactExact time
模型压缩技术（量化、剪枝、知识蒸馏）会不成比例地损害模型的事实准确性，量化后模型在事实性问答任务上的幻觉率可能显著上升
1
Sources
50%
Confidence
Long-term
Relevance
8/6/2026
First Seen
Sources
Related Claims
UnverifiedReasoning capabilities can be compressed through knowledge distillation and reinforcement learning fine-tuning, but the storage of factual knowledge depends on parameter scale, and compression leads to reduced knowledge coverage and higher hallucination rates.76% similarUnverifiedToken compression involves trade-offs: some seemingly redundant contextual information may be critical for model reasoning, and aggressive compression may affect output quality in scenarios such as code generation and complex reasoning.75% similarUnverifiedThe true difficulty of model fine-tuning lies in the construction and cleaning of high-quality training data, preventing catastrophic forgetting, and inference deployment after quantization.74% similarUnverifiedLarge models output the most likely next word rather than precise retrieval of facts, which leads to hallucinations in domains with sparse coverage in the training data.74% similarUnverified过度压缩会损害模型输出质量，压缩效果通常通过困惑度和MMLU、HellaSwag等下游任务准确率衡量73% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/695811API
curl https://kongchang.com/api/v1/knowledge/claims/695811MCP
get_claim(id=695811)