Xeno-Interpretability: Investigating the Alien Minds of LLMs
cs.CL, cs.AI
Submitted: 2026-09-17
Updated: 2026-09-21
License: http://creativecommons.org/licenses/by/4.0/
The gist: Large language models are usually interpreted through concepts that humans already possess: truthfulness, refusal, deception, personality, harmfulness, and related categories.
Terminology
Abstract
Large language models are usually interpreted through concepts that humans already possess: truthfulness, refusal, deception, personality, harmfulness, and related categories. This paper asks whether models may also represent and use distinctions for which no adequate human concept exists. We call such internal structures xeno-representations, and their study xeno-interpretability. We distinguish the human-interpretable semantic space from the xeno-semantic space: the region of model-native representations for which no adequate human conceptual counterpart is available. We show that the space of possible internal distinctions in an LLM is substantially larger than the space available through finite human descriptions. We then separate experimental identification from semantic interpretation: an internal representation may be reproducibly located, geometrically characterized, causally manipulated, and linked to downstream behaviour even when its semantic content cannot be adequately expressed in human terms. On this basis, we sketch an empirical programme to identify xeno-representations. We finally examine the implications for AI safety and multi-agent systems, where model-native representations may propagate and stabilize across interacting agents while remaining only partially visible through human-readable communication. Xeno-interpretability therefore shifts the aim of interpretability from finding human concepts inside models toward discovering and characterizing the representational structures that are native to the models themselves and might affect their behaviour in unpredictable ways.
Sources
- Remarks on the disproof of the unit distance conjecture
- The Alien Space of Science: Sampling Coherent but Cognitively Unavailable Research Directions
- Do Sparse Autoencoders Capture Concept Manifolds?
- Beyond Single-Agent Safety: A Taxonomy of Risks in LLM-to-LLM Interactions
- Institutional AI: Governing LLM Collusion in Multi-Agent Cournot Markets via Public Governance Graphs
- Draining the Energy Commons: Self-Defeating Over-Appropriation as a Coordination Failure in Agentic LLM Collectives
- Persona Vectors: Monitoring and Controlling Character Traits in Language Models
- Not All Language Model Features Are One-Dimensionally Linear
- Mathematical exploration and discovery at scale
- Alignment faking in large language models
- Revisiting the Platonic Representation Hypothesis: An Aristotelian View
- Multi-Agent Risks from Advanced AI
- Mathematical Models of Computation in Superposition
- Sleeper Agents: Training Deceptive LLMs that Persist Through Safety Training
- FinchGPT: a Transformer based language model for birdsong analysis
- Power-seeking can be probable and predictive for trained agents
- The Origins of Representation Manifolds in Large Language Models
- Conversable Complexity: Agentic LLM Collectives as Interpretable Substrates
- Evaluation Awareness Is Not One Capability: Evidence from Open Language Models
Related papers
- Exploring Solution Divergence and Its Effect on Large Language Model Problem Solving
- Ishigaki-IDS-Bench: A Benchmark for Generating Information Delivery Specification from BIM Information Requirements
- Subliminal Steering: Stronger Encoding of Hidden Signals
- MedStruct-S: A Benchmark for Key Discovery, Key-Conditioned QA and Semi-Structured Extraction from OCR Clinical Reports
- The End of Transformers? On Challenging Attention and the Rise of Sub-Quadratic Architectures
- Untangling the Mechanisms of Misleading Context in Medical Question Answering