Unverified50% confidenceConceptExact time
When an Agent can only obtain fragments rather than the full output, cross-fragment reasoning requires multiple retrievals to piece together, and imprecise retrieval causes the model to fill in blanks with 'plausible-looking' content, manifesting as increased hallucinations, and errors accumulate.
1
Sources
50%
Confidence
Long-term
Relevance
10/14/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedContext contamination exists in multi-step agent reasoning, where early erroneous reasoning or redundant information continuously affects subsequent judgments.78% similarUnverified模型规模越大，幻觉的表述往往越流畅、越具迷惑性，识别难度反而上升76% similarUnverifiedWhen faced with ambiguous instructions, large models tend to fill information gaps with content that appears plausible but is unverified (hallucination completion), and explicitly writing out constraints and acceptance criteria can suppress this completion impulse.74% similarUnverifiedEach context summarization and compression performed by an Agent is a lossy encoding, and after multiple rounds of compression, errors cascade and amplify, leading to severe information distortion.74% similarUnverifiedWhen models possess complex multi-step reasoning capabilities, reviewing output content after the fact is no longer sufficient; finer-grained monitoring and intervention at the level of the reasoning process are needed.73% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1024500API
curl https://kongchang.com/api/v1/knowledge/claims/1024500MCP
get_claim(id=1024500)