Unverified50% confidenceCautionExact time
Even when the model is explicitly required to answer only based on retrieved documents, it still fills in gaps using parametric memory and generates seemingly well-formatted citations.
1
Sources
50%
Confidence
Long-term
Relevance
10/1/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedGrounding prompts that instruct the model to answer only based on provided documents and cite the relevant passages can greatly reduce the probability of fabrication, but cannot eliminate it entirely.73% similarUnverifiedThe approach of having the model assemble requests on its own solely based on API documentation has limitations: it may hallucinate parameters, overlook authentication, handle errors improperly, and requires stuffing complete documentation into the context, causing context bloat, increasing costs, and reducing accuracy.72% similarUnverifiedModels are generally lazy and will not proactively save information unless explicitly instructed to do so, requiring users to repeatedly remind them to save.72% similarUnverifiedWhen the model sees an answer key in the folder, it judges that this may be a test trap and proactively chooses not to look at the answers.70% similarUnverifiedLossless compression conclusions reported on contaminated benchmarks are meaningless, because the model can directly retrieve answers from parametric memory rather than relying on attention to retrieve from context.70% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/965583API
curl https://kongchang.com/api/v1/knowledge/claims/965583MCP
get_claim(id=965583)