AIPsy-Affect: A Keyword-Free Clinical Stimulus Battery for Mechanistic Interpretability of Emotion in Language Models

• URL: http://arxiv.org/abs/2604.23719v2
• Date: Tue, 28 Apr 2026 10:21:12 GMT
• Title: AIPsy-Affect: A Keyword-Free Clinical Stimulus Battery for Mechanistic Interpretability of Emotion in Language Models
• Abstract summary: AIPsy-Affect is a 480-item clinical stimulus battery that removes the confound at the stimulus level.<n>It supports linear probing, activation patching, SAE feature analysis, causal ablation, and steering vector extraction.
• Score: 0.0
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Mechanistic interpretability research on emotion in large language models -- linear probing, activation patching, sparse autoencoder (SAE) feature analysis, causal ablation, steering vector extraction -- depends on stimuli that contain the words for the emotions they test. When a probe fires on "I am furious", it is unclear whether the model has detected anger or detected the word "furious". The two readings have very different consequences for every downstream claim about emotion circuits, features, and interventions. We release AIPsy-Affect, a 480-item clinical stimulus battery that removes the confound at the stimulus level: 192 keyword-free vignettes evoking each of Plutchik's eight primary emotions through narrative situation alone, 192 matched neutral controls that share characters, setting, length, and surface structure with the affect surgically removed, plus moderate-intensity and discriminant-validity splits. The matched-pair structure supports linear probing, activation patching, SAE feature analysis, causal ablation, and steering vector extraction under a strong methodological guarantee: any internal representation that distinguishes a clinical item from its matched neutral cannot be doing so on the basis of emotion-keyword presence. A three-method NLP defense battery -- bag-of-words sentiment, an emotion-category lexicon, and a contextual transformer classifier -- confirms the property: bag-of-words methods see only situational vocabulary, and a contextual classifier detects affect (p < 10^-15) but cannot identify the category (5.2% top-1 vs. 82.5% on a keyword-rich control). AIPsy-Affect extends our earlier 96-item battery (arXiv:2603.22295) by a factor of four and is released openly under MIT license.

Related papers

• When Emotion Becomes Trigger: Emotion-style dynamic Backdoor Attack Parasitising Large Language Models [20.883158714880327]
  We propose a pparasitic emotion-style dynamic backdoor attack, Paraesthesia.<n>Paraesthesia achieves an attack success rate of around 99% across both task types and four different models.
  arXiv  Detail & Related papers  (2026-05-12T06:42:36Z)
• Functional Emotions or Situational Contexts? A Discriminating Test from the Mythos Preview System Card [0.0]
  The Claude Mythos Preview system card deploys emotion vectors, sparse autoencoder features, and activation verbalisers to study model internals during misaligned behaviour.<n>This note identifies two hypotheses that are qualitatively consistent with the published results.<n>Which hypothesis is correct determines whether emotion-based monitoring will robustly detect dangerous model behaviour or systematically miss it.
  arXiv  Detail & Related papers  (2026-04-09T19:32:44Z)
• Whether, Not Which: Mechanistic Interpretability Reveals Dissociable Affect Reception and Emotion Categorization in LLMs [0.0]
  We present the first clinical validity test of emotion circuit claims using mechanistic interpretability methods grounded in clinical psychology.<n>We discover two dissociable emotion processing mechanisms.<n>We introduce clinical stimulus methodology as a rigorous standard for testing emotion processing claims in large language models.
  arXiv  Detail & Related papers  (2026-03-15T15:11:45Z)
• Do LLMs "Feel"? Emotion Circuits Discovery and Control [54.57583855608979]
  We study the internal mechanisms that give rise to emotional expression and in controlling emotions in generated text.<n>This is the first systematic study to uncover and validate emotion circuits in large language models.
  arXiv  Detail & Related papers  (2025-10-13T12:24:24Z)
• SEER: The Span-based Emotion Evidence Retrieval Benchmark [8.124633573706761]
  We introduce the SEER (Span-based Emotion Evidence Retrieval) Benchmark to test Large Language Models' ability to identify the specific spans of text that express emotion.<n>We evaluate 14 open-source LLMs and find that, while some models approach average human performance on single-sentence inputs, their accuracy degrades in longer passages.
  arXiv  Detail & Related papers  (2025-10-03T20:15:24Z)
• DiEmo-TTS: Disentangled Emotion Representations via Self-Supervised Distillation for Cross-Speaker Emotion Transfer in Text-to-Speech [49.128847336227636]
  Cross-speaker emotion transfer in speech synthesis relies on extracting speaker-independent emotion embeddings for accurate emotion modeling.<n>We propose DiEmo-TTS, a self-supervised distillation method to minimize emotional information loss and preserve speaker identity.
  arXiv  Detail & Related papers  (2025-05-26T08:47:39Z)
• Dynamic Causal Disentanglement Model for Dialogue Emotion Detection [77.96255121683011]
  We propose a Dynamic Causal Disentanglement Model based on hidden variable separation. This model effectively decomposes the content of dialogues and investigates the temporal accumulation of emotions. Specifically, we propose a dynamic temporal disentanglement model to infer the propagation of utterances and hidden variables.
  arXiv  Detail & Related papers  (2023-09-13T12:58:09Z)
• Unsupervised Extractive Summarization of Emotion Triggers [56.50078267340738]
  We develop new unsupervised learning models that can jointly detect emotions and summarize their triggers. Our best approach, entitled Emotion-Aware Pagerank, incorporates emotion information from external sources combined with a language understanding module.
  arXiv  Detail & Related papers  (2023-06-02T11:07:13Z)
• The MuSe 2022 Multimodal Sentiment Analysis Challenge: Humor, Emotional Reactions, and Stress [71.06453250061489]
  The Multimodal Sentiment Analysis Challenge (MuSe) 2022 is dedicated to multimodal sentiment and emotion recognition. For this year's challenge, we feature three datasets: (i) the Passau Spontaneous Football Coach Humor dataset that contains audio-visual recordings of German football coaches, labelled for the presence of humour; (ii) the Hume-Reaction dataset in which reactions of individuals to emotional stimuli have been annotated with respect to seven emotional expression intensities; and (iii) the Ulm-Trier Social Stress Test dataset comprising of audio-visual data labelled with continuous emotion values of people in stressful dispositions.
  arXiv  Detail & Related papers  (2022-06-23T13:34:33Z)
• SpanEmo: Casting Multi-label Emotion Classification as Span-prediction [15.41237087996244]
  We propose a new model "SpanEmo" casting multi-label emotion classification as span-prediction. We introduce a loss function focused on modelling multiple co-existing emotions in the input sentence. Experiments performed on the SemEval2018 multi-label emotion data over three language sets demonstrate our method's effectiveness.
  arXiv  Detail & Related papers  (2021-01-25T12:11:04Z)
• A Deep Neural Framework for Contextual Affect Detection [51.378225388679425]
  A short and simple text carrying no emotion can represent some strong emotions when reading along with its context. We propose a Contextual Affect Detection framework which learns the inter-dependence of words in a sentence.
  arXiv  Detail & Related papers  (2020-01-28T05:03:15Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.