Task Ecologies and the Evolution of World-Tracking Representations in Large Language Models

• URL: http://arxiv.org/abs/2604.05469v1
• Date: Tue, 07 Apr 2026 06:06:49 GMT
• Title: Task Ecologies and the Evolution of World-Tracking Representations in Large Language Models
• Abstract summary: We study language models as evolving model organisms and ask when autoregressive next-token learning selects for world-tracking representations.<n>For any encoding of latent world states, the Bayes-optimal next-token cross-entropy decomposes into the irreducible conditional entropy plus a Jensen--Shannon excess term.<n>This yields a precise notion of ecological veridicality for language models.
• Score: 0.0
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: We study language models as evolving model organisms and ask when autoregressive next-token learning selects for world-tracking representations. For any encoding of latent world states, the Bayes-optimal next-token cross-entropy decomposes into the irreducible conditional entropy plus a Jensen--Shannon excess term. That excess vanishes if and only if the encoding preserves the training ecology's equivalence classes. This yields a precise notion of ecological veridicality for language models and identifies the minimum-complexity zero-excess solution as the quotient partition by training equivalence. We then determine when this fixed-encoding analysis applies to transformer families: frozen dense and frozen Mixture-of-Experts transformers satisfy it, in-context learning does not enlarge the model's separation set, and per-task adaptation breaks the premise. The framework predicts two characteristic failure modes: simplicity pressure preferentially removes low-gain distinctions, and training-optimal models can still incur positive excess on deployment ecologies that refine the training ecology. A conditional dynamic extension shows how inter-model selection and post-training can recover such gap distinctions under explicit heredity, variation, and selection assumptions. Exact finite-ecology checks and controlled microgpt experiments validate the static decomposition, split-merge threshold, off-ecology failure pattern, and two-ecology rescue mechanism in a regime where the relevant quantities are directly observable. The goal is not to model frontier systems at scale, but to use small language models as laboratory organisms for theory about representational selection.

Related papers

• When the Feature Pool Goes Algorithmic: Extending Mufwene's Ecology of Language Evolution to LLM-Mediated Exposure [0.0]
  Mufwene's ecological model locates language evolution in competition among variants contributed by individual idiolects.<n>Large language models (LLMs) complicate this architecture without requiring the locus of selection to move away from human speakers.<n>I call the resulting ecological process algorithmic reweighting of the speaker-accessible distribution.
  arXiv  Detail & Related papers  (2026-08-21T13:32:26Z)
• Rethinking Entropy Minimization in Test-Time Adaptation for Autoregressive Models [52.18157855348584]
  Test-Time Adaptation (TTA) via entropy minimization (EM) has proven effective for classification tasks, yet its application remains theoretically fragmented.<n>In this work, we deriving a rigorous formulation of EM tailored to autoregressive models.<n>We show that the exact objective naturally decomposes into a token-level policy gradient loss and a token-level entropy loss, and we reinterpret prior methods as partial realizations of this unified formulation.<n>Using Whisper ASR as a testbed, we demonstrate that our approach consistently improves performance across more than 20 diverse domains, including acoustic noise, accents, and multilingual settings.
  arXiv  Detail & Related papers  (2026-05-05T12:00:06Z)
• From Pheromones to Policies: Reinforcement Learning for Engineered Biological Swarms [0.0]
  This study establishes a theoretical equivalence between pheromone-mediated aggregation in celeg and reinforcement learning (RL)<n>We model engineered nematode swarms performing foraging tasks, showing that pheromone dynamics mathematically mirror cross-learning updates.<n>Our results demonstrate that stigmergic systems inherently encode distributed RL processes, where environmental signals act as external memory for collective credit assignment.
  arXiv  Detail & Related papers  (2025-09-24T13:16:35Z)
• Poisson Hierarchical Indian Buffet Processes-With Indications for Microbiome Species Sampling Models [11.64027121881932]
  We introduce a new class of species sampling models designed to address the challenges of complex, sparse count data.<n>Our work provides a broadly applicable methodology for hierarchical count modeling in genetics, commerce, and text analysis.
  arXiv  Detail & Related papers  (2025-02-04T01:27:16Z)
• Principled model selection for stochastic dynamics [0.0]
  We show that overfitting arises not just from individual model complexity but also from the growth of possible models.<n>We introduce Parsimonious Inference (PASTIS), a principled method combining likelihood-estimation with extreme value theory to suppress superfluous parameters.
  arXiv  Detail & Related papers  (2025-01-17T18:23:16Z)
• A Pseudo-Semantic Loss for Autoregressive Models with Logical Constraints [87.08677547257733]
  Neuro-symbolic AI bridges the gap between purely symbolic and neural approaches to learning. We show how to maximize the likelihood of a symbolic constraint w.r.t the neural network's output distribution. We also evaluate our approach on Sudoku and shortest-path prediction cast as autoregressive generation.
  arXiv  Detail & Related papers  (2023-12-06T20:58:07Z)
• DIFFormer: Scalable (Graph) Transformers Induced by Energy Constrained Diffusion [66.21290235237808]
  We introduce an energy constrained diffusion model which encodes a batch of instances from a dataset into evolutionary states. We provide rigorous theory that implies closed-form optimal estimates for the pairwise diffusion strength among arbitrary instance pairs. Experiments highlight the wide applicability of our model as a general-purpose encoder backbone with superior performance in various tasks.
  arXiv  Detail & Related papers  (2023-01-23T15:18:54Z)
• Bayesian Learning of Coupled Biogeochemical-Physical Models [28.269731698116257]
  Predictive models for marine ecosystems are used for a variety of needs. Due to sparse measurements and limited understanding of the myriad of ocean processes, there is significant uncertainty. We develop a Bayesian model learning methodology that allows handling in the space of candidate models and discovery of new models.
  arXiv  Detail & Related papers  (2022-11-12T17:49:18Z)
• Robust Finite Mixture Regression for Heterogeneous Targets [70.19798470463378]
  We propose an FMR model that finds sample clusters and jointly models multiple incomplete mixed-type targets simultaneously. We provide non-asymptotic oracle performance bounds for our model under a high-dimensional learning framework. The results show that our model can achieve state-of-the-art performance.
  arXiv  Detail & Related papers  (2020-10-12T03:27:07Z)
• Goal-directed Generation of Discrete Structures with Conditional Generative Models [85.51463588099556]
  We introduce a novel approach to directly optimize a reinforcement learning objective, maximizing an expected reward. We test our methodology on two tasks: generating molecules with user-defined properties and identifying short python expressions which evaluate to a given target value.
  arXiv  Detail & Related papers  (2020-10-05T20:03:13Z)
• Identification of Probability weighted ARX models with arbitrary domains [75.91002178647165]
  PieceWise Affine models guarantees universal approximation, local linearity and equivalence to other classes of hybrid system. In this work, we focus on the identification of PieceWise Auto Regressive with eXogenous input models with arbitrary regions (NPWARX) The architecture is conceived following the Mixture of Expert concept, developed within the machine learning field.
  arXiv  Detail & Related papers  (2020-09-29T12:50:33Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.