Research
Compositional Generalization via Structural Identification in a Category-Theoretic Framework
arXiv:2608.26465v1 Announce Type: new Abstract: Compositional generalization is usually evaluated through model accuracy. We instead ask which structural or lexical identifications make held-out COGS
arXiv:2608.26465v1 Announce Type: new Abstract: Compositional generalization is usually evaluated through model accuracy. We instead ask which structural or lexical identifications make held-out COGS examples admissible from the structures observed in training. Sentences are represented as functors from syntactic addresses to lexical tokens, and selective collapses induce Kan extensions that propagate observed associations. Across 21 COGS generalization types, admissibility follows distinct identification profiles, while residual failures separate unsupported structural templates. These data-side diagnoses characterize what the training corpus licenses under specified identifications, without training a predictive model.
Related
- ThinkProbe: Beyond Accuracy -- Structural Profiling of Open-Ended LLM Reasoning Traces via Non-Generative Thought Graphs
- How Much Do Circuits Tell Us? Measuring the Consistency and Specificity of Language Model Circuits
- Probing Minimalist Phase Structure in LLMs: What Universal Dependencies Cannot Represent
- CritICL: Inference-Time Weak-to-Strong Generalization from Small Language Model Failure Modes
Source: arXiv cs.CL | 2026-08-28