arXiv:2601. 05280v3 Announce Type: replace-cross Abstract: On the one hand, the question of whether large language models (LLMs) are Solomonoff induction estimators has become an explicit question at the intersection of Algorithmic Information Theory (AIT) and Machine Learning (ML) of great interest.
The paper introduces F-ICL, a benchmark that measures in‑context algorithmic reasoning in language models by exhaustively enumerating 86 million valid programs of length ≤13 on a Turing‑complete machine and computing the exact posterior under a bounded Levin–Solomonoff prior. Unlike typical benchmarks, F‑ICL provides a distributional reference rather than just answers, allowing the evaluation of models’ inductive priors. Across 105 configurations of models ranging from 0.8 B to 675 B parameters, models achieve up to 92 % accuracy, yet many still deviate from the Bayes‑optimal reference, and the study derives theoretical bounds on cumulative loss for predictors with positive prior weight on the reference.
The paper argues that Large Language Models (LLMs) do not function as Solomonoff induction estimators because their training objectives—cross‑entropy, negative log‑likelihood, and next‑token prediction—optimize fit to a supplied conditional distribution rather than a program‑weighted universal mixture. It further contends that additional computation alone does not transform these models into optimal predictors without external hyper‑parameter or architectural changes. The authors suggest that neurosymbolic machine learning, exemplified by models such as Fable and Astra, represents a shift toward symbolic model synthesis, moving beyond purely statistical LLMs.
arXiv:2601.05280v4 Announce Type: replace-cross Abstract: On the one hand, the question of whether Large Language Models (LLMs) are Solomonoff induction estimators has become an explicit question at...
arXiv:2608. 01005v1 Announce Type: new Abstract: Solomonoff Induction, or SolInd, provides an ideal unbounded model of a priori sequence prediction but cannot naturally describe extrapolation from a given training dataset, as performed by Large Language Models.
arXiv:2610.29289v1 Announce Type: new Abstract: Why large language models trained via log-loss minimization over massive and diverse datasets exhibit broad capabilities and generalizable skills is no...
Whether large language models perform genuine algorithmic reasoning or mere pattern completion is hard to test, because most benchmarks lack a ground truth for correct inductive inference. We introduce F-ICL, an in-context-learning benchmark that supplies one exactly.
arXiv:2608. 01575v1 Announce Type: new Abstract: Whether large language models perform genuine algorithmic reasoning or mere pattern completion is hard to test, because most benchmarks lack a ground truth for correct inductive inference.
arXiv:2608. 15798v1 Announce Type: new Abstract: Language models are compared by their held-out per-token cross-entropy risk---the quantity scaling laws are fitted to.
The paper investigates how to extract the key decision points—termed "cruxes"—that influence a language model’s recommendation. Three methods are compared: asking the model to predict its cruxes beforehand, extracting them from a single generated response, and aggregating cruxes across many responses using a reranker. Experiments show that only the many‑rollout approach reliably identifies the full set of important cruxes, making it essential for high‑stakes decisions.
arXiv:2608. 02412v1 Announce Type: new Abstract: Large language models (LLMs) have become the default tool for a remarkable range of tasks, yet they have had conspicuously little success at one of the most common machine learning workloads: predictive analytics over tabular data.
arXiv:2606. 07623v1 Announce Type: new Abstract: This paper develops a model-theoretic framework for verifying context-conditioned language-model behavior by replacing benchmark labels with finite semantic certificates.