What Fits (Into Few Tokens) Doesn't Overfit: Compression and Generalization in ML Research Agents

• URL: http://arxiv.org/abs/2606.11045v1
• Date: Tue, 09 Jun 2026 16:12:55 GMT
• Title: What Fits (Into Few Tokens) Doesn't Overfit: Compression and Generalization in ML Research Agents
• Abstract summary: An attractive hypothesis is that successful machine learning strategies are highly compressible.<n>We study this in the setting of LLM-driven research agents.<n>We find that these bottlenecks have little effect on performance.
• Score: 26.891854386652266
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Reusing a held-out benchmark adaptively should, in principle, invite overfitting. Yet benchmark-driven machine learning (ML) has produced surprisingly little overfitting in practice. An attractive hypothesis is that successful ML strategies are highly compressible. We study this in the setting of LLM-driven research agents, where the hypothesis becomes directly testable via two complementary information bottlenecks. In \emph{output compression}, an exploration agent adaptively searches for high-performance models using a validation set, and we test whether a fresh ``reproducer agent'' can reproduce its performance given only an extremely short prompt and the training data. In \emph{input compression}, the explorer receives only one-bit feedback indicating whether each submitted model improves on the running best. Across 8 datasets spanning tabular classification, vision, language modeling, diffusion modeling, and reward modeling, we find that these bottlenecks have little effect on performance: short prompts and compressible feedback are sufficient to reproduce and find high-performance models. The hypothesis is falsifiable: when we deliberately induce validation-set overfitting, the results fail to reproduce with short prompts. Taken together, our results support a description-length explanation for the lack of overfitting in benchmark-driven ML: successful strategies occupy a low-complexity region of strategy space.

Related papers

• LiveFMBench: Unveiling the Power and Limits of Agentic Workflows in Specification Generation [75.05397479715576]
  Large language models (LLMs) and agents have shown promising progress, but their true capabilities and failure modes remain unclear.<n>We present the first systematic and contamination-aware study of LLM- and agent-based formal specification generation for C programs.
  arXiv  Detail & Related papers  (2026-05-02T11:31:33Z)
• Fine-tuning MLLMs Without Forgetting Is Easier Than You Think [72.59321247529975]
  We design a 2x2 experimental framework to assess model performance across in-distribution and out-of-distribution image and text inputs.<n>Our results show that appropriate regularization, such as constraining the number of trainable parameters or adopting a low learning rate, effectively prevents forgetting when dealing with out-of-distribution images.<n>We attribute this forgetting as task-specific overfitting and address this issue by introducing a data-hybrid training strategy.
  arXiv  Detail & Related papers  (2026-03-15T17:16:19Z)
• Benchmarking Few-shot Transferability of Pre-trained Models with Improved Evaluation Protocols [123.73663884421272]
  Few-shot transfer has been revolutionized by stronger pre-trained models and improved adaptation algorithms.<n>We establish FEWTRANS, a comprehensive benchmark containing 10 diverse datasets.<n>By releasing FEWTRANS, we aim to provide a rigorous "ruler" to streamline reproducible advances in few-shot transfer learning research.
  arXiv  Detail & Related papers  (2026-02-28T05:41:57Z)
• Agentic Test-Time Scaling for WebAgents [65.5178428849495]
  We present Confidence-Aware Test-Time Scaling (CATTS), which uses vote-derived uncertainty to allocate compute only when decisions are genuinely contentious.<n>CATTS improves performance on WebArena-Lite and GoBrowse by up to 9.1% over React while using up to 2.3x fewer tokens than uniform scaling.
  arXiv  Detail & Related papers  (2026-02-12T18:58:30Z)
• Rethinking LLM-as-a-Judge: Representation-as-a-Judge with Small Language Models via Semantic Capacity Asymmetry [41.26991813225211]
  We investigate whether smaller models can serve as efficient evaluators by leveraging internal representations instead of surface generation.<n>We propose the Semantic Capacity Asymmetry Hypothesis: evaluation requires significantly less semantic capacity than generation.<n>We instantiate this paradigm through INSPECTOR, a probing-based framework that predicts aspect-level evaluation scores from small model representations.
  arXiv  Detail & Related papers  (2026-01-30T05:34:24Z)
• The Sweet Danger of Sugar: Debunking Representation Learning for Encrypted Traffic Classification [3.064166155269814]
  This paper critically reassesses proposals that exploit Representation Learning models to create traffic representations.<n>We introduce Pcap-Encoder, an LM-based representation learning model that we specifically design to extract features from protocol headers.<n>Our findings reveal flaws in dataset preparation and model training, calling for a better and more conscious test design.
  arXiv  Detail & Related papers  (2025-07-22T10:32:50Z)
• Rethinking the Role of Prompting Strategies in LLM Test-Time Scaling: A Perspective of Probability Theory [79.63672515243765]
  In this paper, we focus on a standard and realistic scaling setting: majority voting.<n>We show that as the sampling time and computational overhead increase, complicated prompting strategies with superior initial performance gradually fall behind simple Chain-of-Thought.<n>We propose a probabilistic method to efficiently predict scaling performance and identify the best prompting strategy under large sampling times.
  arXiv  Detail & Related papers  (2025-05-16T08:28:57Z)
• Surgical Feature-Space Decomposition of LLMs: Why, When and How? [8.826164604720738]
  We empirically study the efficacy of weight and feature space decomposition in transformer-based language models. We show that surgical decomposition provides critical insights into the trade-off between compression and language modelling performance. We extend our investigation to the implications of low-rank approximations on model bias.
  arXiv  Detail & Related papers  (2024-05-17T07:34:03Z)
• Solving Offline Reinforcement Learning with Decision Tree Regression [0.0]
  This study presents a novel approach to addressing offline reinforcement learning problems by reframing them as regression tasks. We introduce two distinct frameworks: return-conditioned and return-weighted decision tree policies. Despite the simplification inherent in this reformulated approach to offline RL, our agents demonstrate performance that is at least on par with the established methods.
  arXiv  Detail & Related papers  (2024-01-21T23:50:46Z)
• Do Compressed LLMs Forget Knowledge? An Experimental Study with Practical Implications [63.29358103217275]
  Large Language Models (LLMs) often leads to reduced performance, especially for knowledge-intensive tasks. We propose two conjectures on the nature of the damage: one is certain knowledge being forgotten (or erased) after compression. We introduce a variant called Inference-time Dynamic Prompting (IDP) that can effectively increase prompt diversity without incurring any inference overhead.
  arXiv  Detail & Related papers  (2023-10-02T03:12:06Z)
• Meta-Learned Confidence for Few-shot Learning [60.6086305523402]
  A popular transductive inference technique for few-shot metric-based approaches, is to update the prototype of each class with the mean of the most confident query examples. We propose to meta-learn the confidence for each query sample, to assign optimal weights to unlabeled queries. We validate our few-shot learning model with meta-learned confidence on four benchmark datasets.
  arXiv  Detail & Related papers  (2020-02-27T10:22:17Z)

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.