arXiv subjects

Amir Globerson

Publications and source records attributed to Amir Globerson.

At least 19 recordsLinked to original sources

arXiv2026

Towards Scalable Measurement of Durable Skills

Durable skills, such as collaboration, creativity and critical thinking, are instrumental to success in the modern workforce. Yet, measuring these skills remains a persistent challenge. Moreover, because what is not measured is often not taught, these skills are often overlooked in mainstream educational curricula. Designing effective assessments for these skills necessitates balancing two often-conflicting requirements: ecological validity and psychometric rigor. On the one hand, the assessment environment should emulate natural real-world human interaction between humans. On the other hand, it should be scalable, controllable and reproducible. Here we argue that LLMs can be used to better capture both of these aims. Concretely, we develop a framework where the subject converses with AI teammates in a way that resembles human-human interaction for authenticity, while also offering the psychometric control required for informative and robust assessment. Importantly, the AI participants not only act as teammates but also, in an "Executive LLM" setup, steer the conversation towards eliciting a high density of observable evidence for skill proficiency. We complement this with an AI evaluator that can be used to measure skill proficiency in such interactions. We evaluate our assessment protocol based on transcripts of interactions of human participants with our AI framework, for multiple durable skills. For the skill of creativity, we further demonstrate the efficacy of an autorater for evaluating complex tasks performed by real students. Our analysis shows that the use of the Executive LLM significantly increases elicited evidence and that LLM-automated scoring of conversations largely agrees with that of expert annotators. This research demonstrates the utility of orchestrated LLMs approaches for measuring complex social and cognitive constructs in a scalable and controllable manner.

cs.HC↗

arXiv2026

LMs as Task-Specific Knowledge Bases: An Interpretability Analysis

Language models (LMs) capture large amounts of factual knowledge applicable to a wide range of tasks, motivating the view of their parameters as a knowledge base. An important property of knowledge bases is that different queries for the same fact return consistent results, drawing on a single source of truth. We investigate whether LMs satisfy this property through behavioral and mechanistic analyses. Our results suggest that they encode knowledge in a task-specific manner. Behaviorally, facts acquired on one task frequently fail to co-emerge on others during training. Parameter localization experiments suggest a mechanistic explanation, revealing distinct parameter subsets underlying different tasks for the same fact. Finally, we show that chain-of-thought reasoning draws part of its effectiveness from engaging task-specific parameters beyond those tied to the evaluation task. Our findings suggest that what the model knows and how it is asked are intertwined in parameter space, undermining the "knowledge base" analogy and carrying implications for the reliability and controllability of factual knowledge in LMs.

cs.CL↗

arXiv2026

CollabEval: Statistically Efficient Collaborative Model Evaluation via Matrix Completion

Evaluating generative AI models is a routine, but resource-intensive, process that is conducted over and over again during the course of model development. In this work, we propose Collaborative Evaluation (CollabEval), a simple, effective, and principled method for exploiting dependencies between historical runs of different models on the same tasks to improve statistical efficiency. Specifically, our approach treats model evaluation as a matrix completion problem over an $M \times N$ matrix of evaluation scores, where $M$ is the total number of models and $N$ is the total number of evaluation prompts. We assume that a subset of these $M$ models are targeted for evaluation. For these target models only a small fraction, $p$, of prompts has been annotated with evaluation scores. Leveraging recent results in prediction-powered inference, we build a low-rank approximation of the score matrix, and use the reconstructed values as control variates in a manner that guarantees unbiased estimates of the true evaluation metric mean, in addition to statistically valid confidence intervals. Empirically, across a wide range of datasets, models, and sparsity levels $p$, 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.

cs.LG↗

arXiv2026

Depth-Width tradeoffs in Algorithmic Reasoning of Graph Tasks with Transformers

Transformers have revolutionized the field of machine learning. In particular, they can be used to solve complex algorithmic problems, including graph-based tasks. In such algorithmic tasks a key question is what is the minimal size of a transformer that can implement the task. Recent work has begun to explore this problem for graph-based tasks, showing that for sub-linear embedding dimension (i.e., model width) logarithmic depth suffices. However, an open question, which we address here, is what happens if width is allowed to grow linearly, while depth is kept fixed. Here we analyze this setting, and provide the surprising result that with linear width, constant depth suffices for solving a host of graph-based problems. This suggests that a moderate increase in width can allow much shallower models, which are advantageous in terms of inference and train time. For other problems, we show that quadratic width is required. Our results demonstrate the complex and intriguing landscape of transformer implementations of graph-based algorithms. We empirically investigate these trade-offs between the relative powers of depth and width and find tasks where wider models have the same accuracy as deep models, while having much faster train and inference time due to parallelizable hardware.

cs.LG↗

arXiv2026

Cost-Aware Learning

We consider the problem of Cost-Aware Learning, where sampling different components of a finite-sum objective incurs different costs. The objective is to reach a target error while minimizing the total cost. We propose Cost-Aware SGD, which uses a distribution based on gradient norms and costs to sample components. We provide a thorough analysis of this algorithm, including cost-improvement bounds over baselines, a characterization of distribution proxy sub-optimality, and a lower bound. We apply our theoretical insights to reinforcement learning with language models, where the computational cost of sequence-level policy gradients varies with length. We find that the advantage magnitude serves as a high-fidelity proxy for gradient norms, and use this to introduce Cost-Aware GRPO. Empirical results on 1.5B, 4B, and 8B LLMs demonstrate that this algorithm significantly reduces the tokens used in policy optimization while matching or exceeding baseline accuracy.

cs.LG↗

arXiv2026

Lost in Tokenization: Fundamental Trade-offs in Graph Tokenization for Transformers

Transformers have become a central architecture for graph learning, but their application to graphs requires first choosing a tokenization: a graph-to-token map that determines which structural information is exposed at the input. In this work, we show that this choice is a fundamental component of transformer expressivity. We examine three tokenizations that serve as building blocks for many existing graph tokenizations: spectral, random-walk, and adjacency tokenizations. We prove that different tokenizations induce distinct depth regimes: the same graph computation may be realizable by a shallow transformer under one tokenization, while requiring substantially larger depth under another. For example, we prove that random-walk tokenization is lossy for any walk length, making it impossible in general to recover the graph from it, and that while spectral tokenization is lossless, it is ill-conditioned for local tasks. We further show that although both random-walk and spectral tokenizations are derived from adjacency information, it is impossible for a limited-depth transformer to convert between tokenization families in general. In particular, we establish lower bounds and impossibility results showing that unfavorable tokenizations may preclude the efficient recovery of more suitable structural representations. Finally, we complement our theory with controlled experiments on synthetic and real-world tasks, validating the predicted separations and showing that different tasks favor different structural views, and combining complementary tokenizations allows the transformer to leverage distinct signals from each representation.

cs.LG↗

arXiv2026

Controllable User Simulation

Using offline datasets to evaluate conversational agents often fails to cover rare scenarios or to support testing new policies. This has motivated the use of controllable user simulators for targeted, counterfactual evaluation, typically implemented by prompting or fine-tuning large language models. In this work, we formalize controllable simulation as a causal inference problem. By bridging natural language evaluation with off-policy evaluation methodology, we show that the standard practice of training simulators via supervised fine-tuning on post-hoc trajectory labels yields a structurally biased model. Specifically, these labels are inextricably coupled to the data-generating behavior policy, injecting a look-ahead bias that breaks causal consistency. Furthermore, we prove that under policy shift this failure causes the variance of evaluation metrics to explode geometrically, a phenomenon we term controllability collapse. To restore causal consistency, we establish theoretical conditions for accurate simulation and propose practical training mitigations: a priori controls, step-wise dynamic controls, and direct policy-conditioned learning. Empirical evaluation confirms that while standard global controls distort conversational distributions and collapse behavioral diversity, our causally grounded simulators eliminate look-ahead bias, preserve natural variance, and exhibit robust zero-shot generalization to unseen agent behaviors.

cs.AI↗

arXiv2026

Multiple-Prediction-Powered Inference

Statistical estimation often involves tradeoffs between expensive, high-quality measurements and a variety of lower-quality proxies. We introduce Multiple-Prediction-Powered Inference (MultiPPI): a general framework for constructing statistically efficient estimates by optimally allocating resources across these diverse data sources. This work provides theoretical guarantees about the minimax optimality, finite-sample performance, and asymptotic normality of the MultiPPI estimator. Through experiments across three diverse large language model (LLM) evaluation scenarios, we show that MultiPPI consistently achieves lower estimation error than existing baselines. This advantage stems from its budget-adaptive allocation strategy, which strategically combines subsets of models by learning their complex cost and correlation structures.

math.ST↗

arXiv2026

When Can Transformers Count to n?

Large language models based on the transformer architecture can solve highly complex tasks, yet their fundamental limitations on simple algorithmic problems remain poorly understood. In this work, we focus on basic counting tasks and investigate how the difficulty of these tasks scales with the transformer embedding dimension, the context length, and the vocabulary size. We reveal a sharp theoretical phase transition governed by the relationship between the embedding dimension and the vocabulary size. When the dimension is at least as large as the vocabulary, transformers can perfectly maintain token counts. However, when the vocabulary exceeds the embedding dimension, the interference between non-orthogonal token representations forces the network weights to scale polynomially. This renders the exact counting algorithm numerically unstable and practically unlearnable. We empirically validate this bottleneck by training transformers from scratch, demonstrating a strict performance drop at the theoretical threshold and catastrophic out of distribution failure when scaling the vocabulary or context length. Furthermore, we show that state-of-the-art pretrained models suffer from similar failure cases. Our work reveals a critical blind spot absent from the current literature regarding the connection among these three parameters, proving that vocabulary size fundamentally dictates the difficulty of counting tasks.

cs.CL↗

arXiv2026

ConvApparel: A Benchmark Dataset and Validation Framework for User Simulators in Conversational Recommenders

The promise of LLM-based user simulators to improve conversational AI is hindered by a critical "realism gap," leading to systems that are optimized for simulated interactions, but may fail to perform well in the real world. We introduce ConvApparel, a new dataset of human-AI conversations designed to address this gap. Its unique dual-agent data collection protocol -- using both "good" and "bad" recommenders -- enables counterfactual validation by capturing a wide spectrum of user experiences, enriched with first-person annotations of user satisfaction. We propose a comprehensive validation framework that combines statistical alignment, a human-likeness score, and counterfactual validation to test for generalization. Our experiments reveal a significant realism gap across all simulators. However, the framework also shows that data-driven simulators outperform a prompted baseline, particularly in counterfactual validation where they adapt more realistically to unseen behaviors, suggesting they embody more robust, if imperfect, user models.

cs.CL↗

arXiv2026

Latent Reasoning with Supervised Thinking States

Reasoning with a chain-of-thought (CoT) enables Large Language Models (LLMs) to solve complex tasks but incurs significant inference costs due to the generation of long rationales. We propose Thinking States, a method that performs reasoning {\em while} the input is processing. Specifically, Thinking States generates sequences of thinking tokens every few input tokens, transforms the thoughts back into embedding space, and adds them to the following input tokens. This has two key advantages. First, it captures the recurrent nature of CoT, but where the thought tokens are generated as input is processing. Second, since the thoughts are represented as tokens, they can be learned from natural language supervision, and using teacher-forcing, which is parallelizable. Empirically, Thinking States outperforms other latent reasoning methods on multiple reasoning tasks, narrowing the gap to CoT on math problems, and matching its performance on 2-Hop QA with improved latency. On state-tracking tasks, we show Thinking States leads to stronger reasoning behavior than CoT, successfully extrapolating to longer sequences than seen during training.

cs.CL↗

arXiv2025

The FACTS Leaderboard: A Comprehensive Benchmark for Large Language Model Factuality

We introduce The FACTS Leaderboard, an online leaderboard suite and associated set of benchmarks that comprehensively evaluates the ability of language models to generate factually accurate text across diverse scenarios. The suite provides a holistic measure of factuality by aggregating the performance of models on four distinct sub-leaderboards: (1) FACTS Multimodal, which measures the factuality of responses to image-based questions; (2) FACTS Parametric, which assesses models' world knowledge by answering closed-book factoid questions from internal parameters; (3) FACTS Search, which evaluates factuality in information-seeking scenarios, where the model must use a search API; and (4) FACTS Grounding (v2), which evaluates whether long-form responses are grounded in provided documents, featuring significantly improved judge models. Each sub-leaderboard employs automated judge models to score model responses, and the final suite score is an average of the four components, designed to provide a robust and balanced assessment of a model's overall factuality. The FACTS Leaderboard Suite will be actively maintained, containing both public and private splits to allow for external participation while guarding its integrity. It can be found at https://www.kaggle.com/benchmarks/google/facts .

cs.CL↗

arXiv2025

Fast Inference via Hierarchical Speculative Decoding

Transformer language models generate text autoregressively, making inference latency proportional to the number of tokens generated. Speculative decoding reduces this latency without sacrificing output quality, by leveraging a small draft model to propose tokens that the larger target model verifies in parallel. In practice, however, there may exist a set of potential draft models- ranging from faster but less inaccurate, to slower yet more reliable. We introduce Hierarchical Speculative Decoding (HSD), an algorithm that stacks these draft models into a hierarchy, where each model proposes tokens, and the next larger model verifies them in a single forward pass, until finally the target model verifies tokens. We derive an expression for the expected latency of any such hierarchy and show that selecting the latency-optimal hierarchy can be done in polynomial time. Empirically, HSD gives up to 1.2x speed-up over the best single-draft baseline, demonstrating the practicality of our algorithm in reducing generation latency beyond previous techniques.

cs.LG↗

arXiv2025

The Interpretable and Effective Graph Neural Additive Networks

Graph Neural Networks (GNNs) have emerged as the predominant approach for learning over graph-structured data. However, most GNNs operate as black-box models and require post-hoc explanations, which may not suffice in high-stakes scenarios where transparency is crucial. In this paper, we present a GNN that is interpretable by design. Our model, Graph Neural Additive Network (GNAN), is a novel extension of the interpretable class of Generalized Additive Models, and can be visualized and fully understood by humans. GNAN is designed to be fully interpretable, offering both global and local explanations at the feature and graph levels through direct visualization of the model. These visualizations describe exactly how the model uses the relationships between the target variable, the features, and the graph. We demonstrate the intelligibility of GNANs in a series of examples on different tasks and datasets. In addition, we show that the accuracy of GNAN is on par with black-box GNNs, making it suitable for critical applications where transparency is essential, alongside high accuracy.

cs.LG↗

arXiv2025

Towards an AI-Augmented Textbook

Textbooks are a cornerstone of education, but they have a fundamental limitation: they are a one-size-fits-all medium. Any new material or alternative representation requires arduous human effort, so that textbooks cannot be adapted in a scalable manner. We present an approach for transforming and augmenting textbooks using generative AI, adding layers of multiple representations and personalization while maintaining content integrity and quality. We refer to the system built with this approach as Learn Your Way. We report pedagogical evaluations of the different transformations and augmentations, and present the results of a a randomized control trial, highlighting the advantages of learning with Learn Your Way over regular textbook usage.

cs.CY↗

arXiv2025

DeciMamba: Exploring the Length Extrapolation Potential of Mamba

Long-range sequence processing poses a significant challenge for Transformers due to their quadratic complexity in input length. A promising alternative is Mamba, which demonstrates high performance and achieves Transformer-level capabilities while requiring substantially fewer computational resources. In this paper we explore the length-generalization capabilities of Mamba, which we find to be relatively limited. Through a series of visualizations and analyses we identify that the limitations arise from a restricted effective receptive field, dictated by the sequence length used during training. To address this constraint, we introduce DeciMamba, a context-extension method specifically designed for Mamba. This mechanism, built on top of a hidden filtering mechanism embedded within the S6 layer, enables the trained model to extrapolate well even without additional training. Empirical experiments over real-world long-range NLP tasks show that DeciMamba can extrapolate to context lengths that are significantly longer than the ones seen during training, while enjoying faster inference.

cs.LG↗

arXiv2025

Towards Invariance to Node Identifiers in Graph Neural Networks

Message-Passing Graph Neural Networks (GNNs) are known to have limited expressive power, due to their message passing structure. One mechanism for circumventing this limitation is to add unique node identifiers (IDs), which break the symmetries that underlie the expressivity limitation. In this work, we highlight a key limitation of the ID framework, and propose an approach for addressing it. We begin by observing that the final output of the GNN should clearly not depend on the specific IDs used. We then show that in practice this does not hold, and thus the learned network does not possess this desired structural property. Such invariance to node IDs may be enforced in several ways, and we discuss their theoretical properties. We then propose a novel regularization method that effectively enforces ID invariance to the network. Extensive evaluations on both real-world and synthetic tasks demonstrate that our approach significantly improves ID invariance and, in turn, often boosts generalization performance.

cs.LG↗

arXiv2025

On the Expressivity of Selective State-Space Layers: A Multivariate Polynomial Approach

Recent advances in efficient sequence modeling have introduced selective state-space layers, a key component of the Mamba architecture, which have demonstrated remarkable success in a wide range of NLP and vision tasks. While Mamba's empirical performance has matched or surpassed SoTA transformers on such diverse benchmarks, the theoretical foundations underlying its powerful representational capabilities remain less explored. In this work, we investigate the expressivity of selective state-space layers using multivariate polynomials, and prove that they surpass linear transformers in expressiveness. Consequently, our findings reveal that Mamba offers superior representational power over linear attention-based models for long sequences, while not sacrificing their generalization. Our theoretical insights are validated by a comprehensive set of empirical experiments on various datasets.

cs.LG↗