Verified65% confidenceFactTime unknown
Irrelevant information occupying context space (context pollution) can significantly degrade the model's performance on core tasks.
3
Sources
65%
Confidence
Long-term
Relevance
6/2/2026
First Seen
Sources
Related Entities
Related Claims
VerifiedContext pollution refers to irrelevant or harmful information occupying the limited context window space of a model, causing the model's attention to be diverted and reasoning quality to degrade.80% similarUnverifiedSub-agents with narrower task scopes require smaller context windows, directly reducing inference costs and lowering error rates caused by context contamination.79% similarUnverifiedReducing the volume of information before it enters the context is more reliable and more cost-effective than relying on the model to ignore redundancy after it has entered.77% similarUnverifiedWhen multiple unrelated tasks share the same context, it leads to context contamination, degrading the quality of model outputs.76% similarUnverifiedBy injecting large amounts of domain expertise into the model, fine-tuning can significantly reduce the hallucination rate in specific scenarios75% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/38020API
curl https://kongchang.com/api/v1/knowledge/claims/38020MCP
get_claim(id=38020)