Unverified85% confidenceFactTime unknown
负向提示词包含过多token时，每个token在交叉注意力计算中分到的权重被稀释，导致单个约束的实际抑制效力大幅下降
1
Sources
85%
Confidence
Long-term
Relevance
6/2/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedHaving too many tokens in a prompt dilutes the attention mechanism's ability to focus, reducing the actual influence of each description.80% similarUnverified单一智能体面临上下文窗口限制和注意力稀释两大瓶颈，任务复杂度超过单次推理能力时输出质量会显著下降72% similarUnverifiedQuantization error has a non-uniform distribution characteristic; the Q/K/V matrices of the attention mechanism and the last few layers are more sensitive to quantization.71% similarUnverifiedMulti-head Latent Attention significantly reduces memory usage and computational overhead during inference by compressing the KV cache into a low-dimensional latent space.71% similarUnverifiedNegative constraints narrow the probability space by explicitly excluding undesired outputs, are equally important as positive instructions, and can reduce deviations caused by the model's creative improvisation.71% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/36848API
curl https://kongchang.com/api/v1/knowledge/claims/36848MCP
get_claim(id=36848)