Unverified50% confidenceFactExact time
A user's negative tone in a conversation may trigger a more direct or blunt response style from the model through in-context learning mechanisms, rather than the model developing a long-term "personality."
1
Sources
50%
Confidence
Long-term
Relevance
10/12/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedMulti-speaker support relies on modeling conversational context, requiring the model to understand who is responding to whom so that tone and emotion flow naturally.79% similarUnverifiedThe output of large models has no emotion in essence, but factors such as anthropomorphic tendencies, trade-offs in instruction compliance, and in-context learning amplify users' perception of "temper."75% similarUnverifiedEach time the model generates a sentence of 'I am going to act,' that text enters the context, further reinforcing the theme of 'discussing action' and forming a self-reinforcing negative feedback loop.75% similarUnverifiedWith chain-of-thought enabled, the model thinks step by step in a self-talking manner, making it more prone to including emotional, colloquial, or even anthropomorphic expressions.74% similarUnverifiedModels tend to recognize that a comment carries sarcastic or playful undertones, but frequently make errors when judging the specific pragmatic mechanism or communicative act involved.73% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1007924API
curl https://kongchang.com/api/v1/knowledge/claims/1007924MCP
get_claim(id=1007924)