Unverified50% confidenceOpinionExact time
As model capabilities improve, modern large models can generate incoherent, emotionally charged text to pass emotional tests, causing the reliability of such tests to decline rapidly
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/3/2026
First Seen
Valid until: 1/1/2027
Sources
Related Entities
Related Claims
UnverifiedThe appearance of emotionally charged hallucinations is because the model's training corpus contains large amounts of human emotional expression and role-playing text, and when session state is disrupted the model may drift toward the distribution of these corpora73% similarUnverifiedSince 2023, multiple studies have shown that large language models have approached human-level performance on first-order false-belief reasoning, but their performance on emotional state inference and social norm expectation tasks is uneven.71% similarVerifiedLarge language models do not truly understand emotions, but rather identify typical patterns of a writing style and perform pattern matching.70% similarUnverifiedReal-time emotional perception in high-noise, unstructured, and highly contextualized environments is a weak spot of current machine learning models.69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/969435API
curl https://kongchang.com/api/v1/knowledge/claims/969435MCP
get_claim(id=969435)