mlsanityData to 13 Oct 2026

Functional Emotions or Situational Contexts? A Discriminating Test from the Mythos Preview System Card

arXiv 2604.13466 ↗First posted 9 Apr 2026cs.HC cs.AI cs.CLcited by 2

On the map: Mechanistic interpretability → Activation steering methods

Abstract (arXiv)

The Claude Mythos Preview system card deploys emotion vectors, sparse autoencoder (SAE) features, and activation verbalisers to study model internals during misaligned behaviour. The two primary toolkits are not jointly reported on the most alignment-relevant episodes. This note identifies two hypotheses that are qualitatively consistent with the published results: that the emotion vectors track functional emotions that causally drive behaviour, or that they are a projection of a richer situational-context structure onto human emotional axes. The hypotheses can be distinguished by cross-referencing the two toolkits on episodes where only one is currently reported: most directly, applying emotion probes to the strategic concealment episodes analysed only with SAE features. If emotion probes show flat activation while SAE features are strongly active, the alignment-relevant structure lies outside the emotion subspace. Which hypothesis is correct determines whether emotion-based monitoring will robustly detect dangerous model behaviour or systematically miss it.

Key facts. Cited by 2 papers in the map (data through 13 Oct 2026). Placed in Activation steering methods, one of 13 directions in Mechanistic interpretability. Builds on 5 of its 7 references that are in the map. First posted on arXiv 9 Apr 2026.

Builds on

5 of 7 references.

References · 5

  1. 2021Probing Classifiers: Promises, Shortcomings, and Advancescited by 850
  2. 2019Designing and Interpreting Probes with Control Taskscited by 620
  3. 2025Persona Vectors: Monitoring and Controlling Character Traits in Language Modelscited by 321
  4. 2025Auditing language models for hidden objectivescited by 76
  5. 2026AI Arms and Influence: Frontier Models Exhibit Sophisticated Reasoning in Simulated Nuclear Crisescited by 7

Built on by

Cited by year (2026 to 13 Oct)

Citing papers per year: 2026: 2.

2 papers.

  1. 2026Replicating the Geometry of Emotion Representations in a Base Open-Weights Modelcited by 0
  2. 2026Same Outcome, Different Readout: What Does a Steerable Valence Direction in LLMs Represent?cited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026Why Are Some Emotions Harder for LLMs? Uncovering the Causal Mechanisms of Emotion Inference via Sparse Autoencoders
  2. 2026Emotion Concepts and their Function in a Large Language Model
  3. 2026Cognition-Oriented Emotion Tracing from Causes to Consequences in Real-World Social Scenes
  4. 2025Bridging the behavior-neural gap: A multimodal AI reveals the brain's geometry of emotion more accurately than human self-reports
  5. 2026Faithful Action-unit Causal Reasoning for Counterfactually Faithful Emotion Explanations
  6. 2026How Emotion Shapes the Behavior of LLMs and Agents: A Mechanistic Study
  7. 2025Emotions Where Art Thou: Understanding and Characterizing the Emotional Latent Space of Large Language Models
  8. 2026The Echo Amplifies the Knowledge: Somatic Marker Analogues in Language Models via Emotion Vector Re-Injection