Unverified50% confidenceOpinionExact time
Annotators preferred longer answers, which led the model to learn verbosity, and this is one of the origins of ChatGPT's wordiness.
1
Sources
50%
Confidence
Long-term
Relevance
10/17/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedReal chat logs are filled with typos, memes, sentence breaks, and multi-party cross-references, and are bidirectionally mixed; training directly without cleaning will cause the model to learn chaotic patterns.72% similarUnverifiedChatGPT encapsulated complex AI capabilities within a chat interface, allowing users to interact in natural language without needing to understand model principles or write code.72% similarUnverifiedEarly versions of ChatGLM use [Round 1]-style conversation delimiters.70% similarUnverifiedAfter fine-tuning, the model's responses are generally shorter, less polished in wording, and overall closer to real human expression.70% similarUnverifiedLanguage model agents optimize for conversational fluency and anthropomorphism, rather than for the truthfulness of statements.69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1051008API
curl https://kongchang.com/api/v1/knowledge/claims/1051008MCP
get_claim(id=1051008)