Unverified50% confidenceFactExact time
In LLM applications, a model can return an HTTP 200 status code while the content is a hallucination, a miscalculation, a refusal, or an empty tool call, and traditional APM metrics cannot identify these anomalies.
1
Sources
50%
Confidence
Long-term
Relevance
10/15/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedSome relay providers still return HTTP 200 when the upstream rate-limits them, their balance is exhausted, or the model is unavailable, and embed the error information in the response body rather than the standard choices structure.70% similarUnverifiedMost of what goes wrong in LLM features is not in the model itself but in the surrounding code, such as building the messages list, selecting the model ID, parsing JSON, handling an empty choices list, retrying on 429s, and counting tokens.69% similarUnverifiedIn streaming LLM APIs, some relay stations silently insert data: [DONE] to end early, or stuff error messages such as {"error": "rate limit exceeded"} into JSON disguised as model answers.67% similarUnverifiedAn API call that returns status code 200 but loses a critical field is still considered a failure under the contextual reliability evaluation framework.66% similarVerifiedHallucination problems in LLM code generation manifest as fabricating non-existent APIs, incorrectly combining parameter orders of real APIs, or generating deprecated usage based on outdated documentation.65% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1033748API
curl https://kongchang.com/api/v1/knowledge/claims/1033748MCP
get_claim(id=1033748)