The paper argues that as Large Language Models transition from chatbots to agentic systems, the current post-hoc interpretability paradigm is insufficient for safe deployment because it cannot audit or intervene before an output is produced. It proposes a shift to generative interpretability, where a model’s inference process inherently exposes semantically meaningful checkpoints that are human-understandable and can be causally intervened upon. The authors illustrate the advantages of this approach and introduce Neuro‑Symbolic Models as a concrete implementation.
arXiv:2606. 11195v1 Announce Type: cross Abstract: Large language models (LLMs) have transformed how humans access information, but not how we reason with it.
arXiv:2606. 26523v1 Announce Type: new Abstract: We develop a framework for interpreting AI systems as agents, drawing on the philosophical tradition of radical interpretation and the tools of mechanistic interpretability.
Neural networks are likely to be among the most consequential artefacts in human history. As these systems become more capable and autonomous, their potential to cause societal harm also grows, making...
arXiv:2611.07088v1 Announce Type: new Abstract: Symbolic models offer a powerful account of cognition, but their scope has historically been limited by the need for hand-engineering and computational...
arXiv:2604. 19775v2 Announce Type: replace Abstract: Large Language Models (LLMs) are increasingly deployed as autonomous agents capable of reasoning, planning, and acting within interactive environments.
arXiv:2608. 12325v1 Announce Type: new Abstract: Autonomous reasoning is among the most scientifically and economically motivating topics in AI today.
The paper surveys mechanistic interpretability for AI safety, exploring how studying internal representations and computations of neural networks can help evaluate, predict, and control their behavior. It reviews foundational concepts such as features, polysemanticity, transformer computation, and circuits, and surveys methods for feature extraction, circuit discovery, causal verification, and training-data attribution. The authors propose a unified framework for evaluating mechanistic claims, outline six research orientations, and discuss how mechanistic understanding can aid safety through knowledge editing, steering, and monitoring, while also highlighting limitations and open problems.
arXiv:2607. 07229v1 Announce Type: new Abstract: Prior work has shown that chain-of-thought (CoT) reasoning is often unfaithful: a model's stated reasoning does not reliably reflect the process that produced its output.
The paper introduces the concept of xeno-interpretability, which studies internal distinctions in large language models that lack corresponding human concepts. It distinguishes between human‑interpretable and xeno‑semantic spaces, showing that LLMs possess a far larger internal representational space than can be captured by finite human descriptions. The authors propose an empirical program to identify and characterize these xeno‑representations, noting their potential to influence model behavior in ways that are not fully visible through human‑readable communication.
arXiv:2602. 13904v2 Announce Type: replace Abstract: Chain-of-thought (CoT) reasoning is fundamental to modern LLM architectures and represents a critical intervention point for AI safety.
arXiv:2510. 14538v3 Announce Type: replace Abstract: Neuro-symbolic (NeSy) AI aims to develop deep neural networks whose predictions comply with prior knowledge encoding, e.
The paper proposes a user‑centric Chain‑of‑Thought (CoT) reasoning framework that structures LLM reasoning traces into self‑contained, verifiable steps using XML‑like tags. This design allows users to independently assess and correct the AI’s reasoning while preserving performance on mathematical reasoning tasks. User studies show that the approach improves perceived usefulness and ease of use compared to standard CoT.