arXiv AI
Aug 10
Geo-Spatial Concept Probing of Large Language Models: Abstraction, Compositionality, and Grounding
arXiv:2608. 07353v1 Announce Type: cross Abstract: Understanding concepts is fundamental to generalization.
MAxBench: A Multinomial Concept Recovery Benchmark
MAxBench is a geometry‑agnostic benchmark for evaluating how well language models recover multinomial concept representations. The study compares ten localization methods across five geometry types, six concepts, and four models, finding that affine subspaces generally steer more reliably and recall more instances than rank‑one or linear subspaces. The results also show that manifold steering can match the best methods when applicable, and that no method consistently outperforms prompting for these complex concepts.
Related stories
arXiv AI
3d ago
Interpretable but Fragile? Robustness of Concept Bottlenecks under Geometric-Semantic Perturbations
arXiv:2609.38625v1 Announce Type: cross Abstract: Concept Bottleneck Models (CBMs) are designed to provide interpretable intermediate representations, yet how such bottlenecks affect robustness remai...
arXiv Machine Learning
1d ago
Beyond Linear Concepts: Discovering and Aligning Non-Linear Concept Manifolds in Large Language Models
The paper extends mechanistic interpretability of large language models by modeling concepts as low‑dimensional non‑linear manifolds rather than linear subspaces. It introduces a concept‑based alignment (CBA) score to compare these manifolds across layers and models, revealing block structures in intermediate layers, a shift from syntax‑dominated to mixed syntactic‑semantic concepts, and training‑dependent multilingual sharing. The study also shows that alignment patterns differ across model families and training stages, with adjacent stages aligning more closely than distant ones.
arXiv Computation and Language
Sep 1
The First Token Is a Clue: Verbalizing Multi-Token Concepts from the J-lens
arXiv:2608.31084v1 Announce Type: new Abstract: The Jacobian Lens (J-lens) is a recent tool for interpreting LLMs. It reads a hidden state as a ranked list of vocabulary tokens, leaving multi-token c...
arXiv AI
Jun 9
A Geometric Unification of Concept Learning with Concept Cones
arXiv:2512. 07355v2 Announce Type: replace Abstract: Two traditions of interpretability have evolved side by side but seldom spoken to each other: Concept Bottleneck Models (CBMs), which prescribe what a concept should be, and Sparse Autoencoders (SAEs), which discover what concepts emerge.
arXiv Machine Learning
Jun 9
A Unifying Framework for Concept-Based Representational Similarity
arXiv:2606. 09653v1 Announce Type: new Abstract: Learned representations across models and modalities often exhibit striking structural similarities, suggesting shared underlying concept decompositions.
arXiv AI
Jul 14
Laguerre Geometry for Interpreting Large Language Models
arXiv:2607. 10578v1 Announce Type: new Abstract: Existing hypotheses represent a concept in an LLM as a single point, a linear direction, or a Gaussian cluster, yet it remains unclear how and why such structures emerge.
arXiv AI
Jul 7
Unsupervised Features Mining via Activation Geometry
arXiv:2607. 04222v1 Announce Type: new Abstract: Interpretability methods aim to reveal the features represented inside large language models (LLMs).
arXiv AI
Jul 10
Understanding Axes of Difficulty For Long Context Tasks Via PredicateLongBench
arXiv:2607. 08284v1 Announce Type: new Abstract: Large language models (LLMs) have demonstrated rapidly improving long-context capabilities, prompting a wave of benchmarks designed to evaluate them.
arXiv Computation and Language
Aug 21
ConceptGuard: Benchmarking Context-Sensitive Unlearning in Large Language Models
arXiv:2608. 20338v1 Announce Type: new Abstract: Large Language Models (LLMs) increasingly require selective removal of harmful or sensitive knowledge, called unlearning, yet existing methods and benchmarks fail to evaluate this capability completely.
arXiv AI
Jun 11
ICA Lens: Interpreting Language Models Without Training Another Dictionary
arXiv:2606. 11722v1 Announce Type: cross Abstract: Finding interpretable directions in language-model representations is critical for understanding and controlling model behavior.
arXiv AI
Jun 10
PhantomBench: Benchmarking the Non-existential Threat of Language Models
arXiv:2606. 11105v1 Announce Type: cross Abstract: Hallucinations, where language models (LMs) generate factually ungrounded responses, pose serious risks, as users tend to blindly rely on them.