Researchers challenge emotion-vector paper: 80% of examples may just be negative emotion
rgblong · x · 2026-10-09
In a thread with dcshiller and others, rgblong challenges a paper on pain/emotion vectors: despite claiming novelty over prior emotion concepts work, 80% of its vector examples look like plain negative emotion, blurring the line with claimed 'psychological, social, moral-cognitive pain'. The discussion also questions whether the denoising step (projecting out control-group principal components) removes sub-components like body signals, and whether subtracting absent controls effectively adds their inverse.
Related event: Researchers Challenge the 'Pain Axis' Paper as Author Defends Method(16 posts)→
More from Research
- Are URM and Universal Transformers the forgotten architecture beating standard LLMs? — moschles · 2026-10-09
- Researcher: Use AI to Rewrite Machine-Generated Math Proofs Into Human-Readable Forms — jd_pressman · 2026-10-09
- AutoScientist's two-agent checklist loop auto-audits every training example — sarahookr · 2026-10-09
- NeurIPS GenAI4Health oral: retrieval-based medical fact-checking fails in ways bigger models can't fix — mdredze · 2026-10-09
- Bigger models, more reasoning, better sources won't fix medical fact-checking, researchers say — mdredze · 2026-10-09
- DEX best abstract: top LLMs catch many physician diagnostic errors, but big gaps remain — mdredze · 2026-10-09