Valid Best-Model Identification for LLM Evaluation via Low-Rank Factorization

• URL: http://arxiv.org/abs/2605.10405v1
• Date: Mon, 11 May 2026 11:43:28 GMT
• Title: Valid Best-Model Identification for LLM Evaluation via Low-Rank Factorization
• Abstract summary: We propose a principled framework that combines MAB with cheap predicted scores without compromising statistical validity.<n>We derive doubly robust estimators of each model's performance that use the low-rank predictions to reduce variance.<n> Empirical results on real-world benchmarks show that our approach reduces the number of required evaluations, yielding meaningful savings in compute and cost.
• Score: 17.34147279018477
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Selecting the best large language model (LLM) for a fixed benchmark is often expensive, since exhaustive evaluation requires running every model on every example. Multi-armed bandit (MAB) algorithms can reduce the number of LLM calls by sequentially selecting the next model-example pair to evaluate, thereby avoiding wasted evaluations on clearly underperforming models. Further savings can be achieved by predicting model scores from the partially observed model-example score matrix using low-rank factorization. However, such predictions are not ground truth: they can be biased and may therefore lead to incorrect identification of the best model. In this work, we propose a principled framework that combines MAB with cheap predicted scores without compromising statistical validity. Specifically, we derive doubly robust estimators of each model's performance that use the low-rank predictions to reduce variance. This enables the construction of valid finite-sample confidence intervals in our setting, where models are selected adaptively and examples are sampled without replacement. Empirical results on real-world benchmarks show that our approach reduces the number of required evaluations, yielding meaningful savings in compute and cost while accurately identifying the best-performing model.

Related papers

• Dynamically Allocating Evaluation Effort for Model Ranking [60.00410111594536]
  We formalize multi-model human evaluation as a best-arm identification problem in a multi-armed bandit setup with correlated arms.<n>We show that it improves discrimination between top-performing models.<n>This makes evaluations faster, cheaper and more aligned with large-scale competition evaluation goals.
  arXiv  Detail & Related papers  (2026-08-04T10:33:04Z)
• CollabEval: Statistically Efficient Collaborative Model Evaluation via Matrix Completion [65.53153291902454]
  Collaborative Evaluation (CollabEval) is a simple, effective, and principled method for exploiting dependencies between historical runs of different models on the same tasks to improve statistical efficiency.<n>We build a low-rank approximation of the score matrix, and use the reconstructed values to guarantee unbiased estimates of the true evaluation metric mean.<n>We find that CollabEval substantially reduces the mean confidence interval size, and the mean squared error of the point estimate, compared to baseline methods at the same annotation budget.
  arXiv  Detail & Related papers  (2026-07-06T13:22:04Z)
• Cutting LLM Evaluation Costs with SySRs: A Bandit Algorithm that Provably Exploits Model Similarity [14.830460628644715]
  Large Language Models are typically benchmarked by evaluating every model on every test query.<n>We propose Synchronized Successive Rejects (SySRs), augmenting the classical Successive Rejects algorithm with paired comparisons.
  arXiv  Detail & Related papers  (2026-06-05T17:03:19Z)
• A Theoretical Framework for Statistical Evaluability of Generative Models [57.9316356505791]
  We introduce a theoretical framework for evaluating generative models and establish evaluability results for commonly used metrics.<n>We study two categories of metrics: test-based metrics, including integral probability metrics (IPMs) and Rényi divergences.<n>We show that IPMs with respect to any bounded test class can be evaluated from finite samples up to multiplicative and additive approximation errors.<n>In contrast, Rényi and KL divergences are not evaluable from finite samples, as their values can be critically determined by rare events.
  arXiv  Detail & Related papers  (2026-04-07T01:53:59Z)
• Fault-Tolerant Evaluation for Sample-Efficient Model Performance Estimators [13.227055178509524]
  We propose a fault-tolerant evaluation framework that integrates bias and variance considerations within an adjustable tolerance level.<n>We show that proper calibration of $varepsilon$ ensures reliable evaluation across different variance regimes.<n> Experiments on real-world datasets demonstrate that our framework provides comprehensive and actionable insights into estimator behavior.
  arXiv  Detail & Related papers  (2026-02-06T22:14:46Z)
• From Model Choice to Model Belief: Establishing a New Measure for LLM-Based Research [0.0]
  Large language models (LLMs) are increasingly used to simulate human behavior.<n>Treating an LLM's output as a single data point underutilizes the information inherent to the probabilistic nature of LLMs.<n>This paper introduces and formalizes "model belief," a measure derived from an LLM's token-level probabilities.
  arXiv  Detail & Related papers  (2025-12-29T03:50:40Z)
• Model Correlation Detection via Random Selection Probing [62.093777777813756]
  Existing similarity-based methods require access to model parameters or produce scores without thresholds.<n>We introduce Random Selection Probing (RSP), a hypothesis-testing framework that formulates model correlation detection as a statistical test.<n>RSP produces rigorous p-values that quantify evidence of correlation.
  arXiv  Detail & Related papers  (2025-09-29T01:40:26Z)
• Self-Boost via Optimal Retraining: An Analysis via Approximate Message Passing [58.52119063742121]
  Retraining a model using its own predictions together with the original, potentially noisy labels is a well-known strategy for improving the model performance.<n>This paper addresses the question of how to optimally combine the model's predictions and the provided labels.<n>Our main contribution is the derivation of the Bayes optimal aggregator function to combine the current model's predictions and the given labels.
  arXiv  Detail & Related papers  (2025-05-21T07:16:44Z)
• Ranked from Within: Ranking Large Multimodal Models Without Labels [73.96543593298426]
  We show that uncertainty scores derived from softmax distributions provide a robust basis for ranking models across various tasks.<n>This facilitates the ranking of LMMs on unlabeled data, providing a practical approach for selecting models for diverse target domains without requiring manual annotation.
  arXiv  Detail & Related papers  (2024-12-09T13:05:43Z)
• Post-Selection Confidence Bounds for Prediction Performance [2.28438857884398]
  In machine learning, the selection of a promising model from a potentially large number of competing models and the assessment of its generalization performance are critical tasks. We propose an algorithm how to compute valid lower confidence bounds for multiple models that have been selected based on their prediction performances in the evaluation set.
  arXiv  Detail & Related papers  (2022-10-24T13:28:43Z)
• Model-based metrics: Sample-efficient estimates of predictive model subpopulation performance [11.994417027132807]
  Machine learning models $-$ now commonly developed to screen, diagnose, or predict health conditions are evaluated with a variety of performance metrics. Subpopulation performance metrics are typically computed using only data from that subgroup, resulting in higher variance estimates for smaller groups. We propose using an evaluation model $-$ a model that describes the conditional distribution of the predictive model score $-$ to form model-based metric (MBM) estimates.
  arXiv  Detail & Related papers  (2021-04-25T19:06:34Z)
• Characterizing Fairness Over the Set of Good Models Under Selective Labels [69.64662540443162]
  We develop a framework for characterizing predictive fairness properties over the set of models that deliver similar overall performance. We provide tractable algorithms to compute the range of attainable group-level predictive disparities. We extend our framework to address the empirically relevant challenge of selectively labelled data.
  arXiv  Detail & Related papers  (2021-01-02T02:11:37Z)
• SUMO: Unbiased Estimation of Log Marginal Probability for Latent Variable Models [80.22609163316459]
  We introduce an unbiased estimator of the log marginal likelihood and its gradients for latent variable models based on randomized truncation of infinite series. We show that models trained using our estimator give better test-set likelihoods than a standard importance-sampling based approach for the same average computational cost.
  arXiv  Detail & Related papers  (2020-04-01T11:49:30Z)

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.