LLMs can solve complex tasks by generating long, multi-step reasoning chains. Test-time scaling (TTS) can further improve LLM performance by sampling multiple variants of intermediate reasoning steps, verifying their correctness, and strategically choosing the best steps for continuation. However, existing verification approaches, such as Process Reward Models (PRMs), are computationally expensive, limited to specific domains, and require large-scale human or model-generated annotations. We propose a lightweight alternative for step-level reasoning verification based on probing the internal states of LLMs. We train a transformer-based probe that uses the internal states of the frozen LLM to estimate the credibility of its reasoning steps during generation. Annotation can be generated either by another larger LLM (e.g., DeepSeek-R1) or in a self-supervised manner by the original model itself. The probes are both effective and lightweight, containing fewer than 10M parameters. Across multiple domains, including mathematics, planning, and general knowledge question answering, our probes match or even exceed the performance of PRMs that are up to 810× larger. Our findings suggest that the internal states of LLMs encode their confidence in reasoning processes and can serve as reliable signals for reasoning step verification, offering a promising direction towards scalable and generalizable TTS and introspective LLMs.

Test-time compute (TTC) scaling has emerged as a powerful paradigm for improving large language model (LLM) reasoning by allocating additional compute during inference, e.g., via multi-sample generation and verifier-based reranking. Existing TTC scaling strategies and reasoning scorers remain fragmented, evaluated under inconsistent protocols, and are rarely analyzed through the lens of quality-cost trade-offs. We introduce ThinkBooster, a unified framework for seamless test-time compute scaling of LLM reasoning, which consists of (i) a modular Python library implementing state-of-the-art TTC scaling strategy and scorer families, (ii) a benchmark that jointly evaluates performance and computational efficiency, and (iii) a deployable OpenAI-compatible proxy service that enables drop-in integration of adaptive reasoning into real-world applications. We further provide a demo visual debugger for inspecting the reasoning trajectories, intermediate selection decisions, and alternative reasoning paths. Empirical results on mathematical and coding tasks reveal the performance-compute trade-offs of TTC scaling strategies and scoring methods and demonstrate that ThinkBooster provides practical gains in real-world tasks. The code is available online under an MIT license.

The rapid proliferation of large language models (LLMs) has stimulated researchers to seek effective and efficient approaches to deal with LLM hallucinations and low-quality outputs. Uncertainty quantification (UQ) is a key element of machine learning applications in dealing with such challenges. However, research to date on UQ for LLMs has been fragmented in terms of techniques and evaluation methodologies. In this work, we address this issue by introducing a novel benchmark that implements a collection of state-of-the-art UQ baselines and offers an environment for controllable and consistent evaluation of novel UQ techniques over various text generation tasks. Our benchmark also supports the assessment of confidence normalization methods in terms of their ability to provide interpretable scores. Using our benchmark, we conduct a large-scale empirical investigation of UQ and normalization techniques across eleven tasks, identifying the most effective approaches.

LLMs have the tendency to hallucinate, i.e., to sporadically generate false or fabricated information, and users generally lack the tools to detect when this happens. Uncertainty quantification (UQ) provides a framework for assessing the reliability of model outputs, aiding in the identification of potential hallucinations. In this work, we introduce pre-trained UQ heads: supervised auxiliary modules for LLMs that substantially enhance their ability to capture uncertainty compared to unsupervised UQ methods. Their strong performance stems from the transformer architecture in their design, in the form of informative features derived from LLM attention maps and logits. Our experiments show that these heads are highly robust and achieve state-of-the-art performance in claim-level hallucination detection across both in-domain and out-of-domain prompts. Moreover, these modules demonstrate strong generalization to languages they were not explicitly trained on. We pre-train a collection of UQ heads for popular LLM series, including Mistral, Llama, and Gemma. We publicly release both the code and the pre-trained heads.

Uncertainty quantification (UQ) has emerged as a promising approach for detecting hallucinations and low-quality output of Large Language Models (LLMs). However, obtaining proper uncertainty scores is complicated by the conditional dependency between the generation steps of an autoregressive LLM, because it is hard to model it explicitly. Here, we propose to learn this dependency from attention-based features. In particular, we train a regression model that leverages LLM attention maps, probabilities on the current generation step, and recurrently computed uncertainty scores from previously generated tokens. To incorporate the recurrent features, we also suggest a two-staged training procedure. Our experimental evaluation on ten datasets and three LLMs shows that the proposed method is highly effective for selective generation, achieving substantial improvements over rivaling unsupervised and supervised approaches.

Large language models (LLMs) are widely used in NLP applications, but their tendency to produce hallucinations poses significant challenges to the reliability and safety, ultimately undermining user trust. This tutorial offers the first systematic introduction to uncertainty quantification (UQ) for LLMs in text generation tasks – a conceptual and methodological framework that provides tools for communicating the reliability of a model answer. This additional output could be leveraged for a range of downstream tasks, including hallucination detection and selective generation. We begin with the theoretical foundations of uncertainty, highlighting why techniques developed for classification might fall short in text generation. Building on this grounding, we survey state-of-the-art white-box and black-box UQ methods, from simple entropy-based scores to supervised probes over hidden states and attention weights, and show how they enable selective generation and hallucination detection. Additionally, we discuss the calibration of uncertainty scores for better interpretability. A key feature of the tutorial is practical examples using LM-Polygraph, an open-source framework that unifies more than a dozen recent UQ and calibration algorithms and provides a large-scale benchmark, allowing participants to implement UQ in their applications, as well as reproduce and extend experimental results with only a few lines of code. By the end of the session, researchers and practitioners will be equipped to (i) evaluate and compare existing UQ techniques, (ii) develop new methods, and (iii) implement UQ in their code for deploying safer, more trustworthy LLM-based systems.

Large language models (LLMs) are notorious for hallucinating, i.e., producing erroneous claims in their output. Such hallucinations can be dangerous, as occasional factual inaccuracies in the generated text might be obscured by the rest of the output being generally factually correct, making it extremely hard for the users to spot them. Current services that leverage LLMs usually do not provide any means for detecting unreliable generations. Here, we aim to bridge this gap. In particular, we propose a novel fact-checking and hallucination detection pipeline based on token-level uncertainty quantification. Uncertainty scores leverage information encapsulated in the output of a neural network or its layers to detect unreliable predictions, and we show that they can be used to fact-check the atomic claims in the LLM output. Moreover, we present a novel token-level uncertainty quantification method that removes the impact of uncertainty about what claim to generate on the current step and what surface form to use. Our method Claim Conditioned Probability (CCP) measures only the uncertainty of a particular claim value expressed by the model. Experiments on the task of biography generation demonstrate strong improvements for CCP compared to the baselines for seven different LLMs and four languages. Human evaluation reveals that the fact-checking pipeline based on uncertainty quantification is competitive with a fact-checking tool that leverages external knowledge.

Recent advancements in the capabilities of large language models (LLMs) have paved the way for a myriad of groundbreaking applications in various fields. However, a significant challenge arises as these models often “hallucinate”, i.e., fabricate facts without providing users an apparent means to discern the veracity of their statements. Uncertainty estimation (UE) methods are one path to safer, more responsible, and more effective use of LLMs. However, to date, research on UE methods for LLMs has been focused primarily on theoretical rather than engineering contributions. In this work, we tackle this issue by introducing LM-Polygraph, a framework with implementations of a battery of state-of-the-art UE methods for LLMs in text generation tasks, with unified program interfaces in Python. Additionally, it introduces an extendable benchmark for consistent evaluation of UE techniques by researchers, and a demo web application that enriches the standard chat dialog with confidence scores, empowering end-users to discern unreliable responses. LM-Polygraph is compatible with the most recent LLMs, including BLOOMz, LLaMA-2, ChatGPT, and GPT-4, and is designed to support future releases of similarly-styled LMs.