Unverified50% confidenceFactExact time
魏杰森（Jason Wei）等人2022年的研究发现，当模型规模超过某个临界点，情感推理等任务的表现会发生非线性跃升（涌现能力）
1
Sources
50%
Confidence
Long-term
Relevance
7/13/2026
First Seen
Sources
In-Depth Analysis of the GPT-5.6 EQ Test: Progress and Limitations in Emotional Intelligence
redditr/OpenAI7/11/2026
Related Claims
Unverified当推荐模型的优化目标被设定为最大化engagement时，模型会系统性给予情绪化内容更高分发权重，这被研究者称为算法放大效应（algorithmic amplification）68% similarUnverifiedAn Anthropic co-founder acknowledged that what the emotion-like signals discovered inside AI models mean remains unclear and requires ongoing discernment.65% similarUnverified有研究发现，在提示词中加入情感激励语句（如'这对我的职业生涯很重要'），模型在数学推理任务上的准确率提升约8-15%64% similarUnverifiedAI直接描述情绪在统计上是更"安全"、更高频的表达选择，符合模型对"好写作"的概率预测64% similarUnverifiedAnthropic researchers have discovered emotion-like signals inside AI models63% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/497463API
curl https://kongchang.com/api/v1/knowledge/claims/497463MCP
get_claim(id=497463)