An official website of the United States government
An interesting behavior in large language models (LLMs) is prompt sensitivity. When provided with different but semantically equivalent versions of the same prompt, models may produce very different distributions of answers. This suggests that the uncertainty reflected in a model's output distribution for one prompt may not reflect the model's uncertainty about the meaning of the prompt. We model prompt sensitivity as a type of generalization error, and show that sampling across the semantic concept space with paraphrasing perturbations improves uncertainty calibration without compromising accuracy. Additionally, we introduce a new metric for uncertainty decomposition in black-box LLMs that improves upon entropy-based decomposition by modeling semantic continuities in natural language generation. We show that this decomposition metric can be used to quantify how much LLM uncertainty is attributed to prompt sensitivity. Our work introduces a new way to improve uncertainty calibration in prompt-sensitive language models, and provides evidence that some LLMs fail to exhibit consistent general reasoning about the meanings of their inputs.
more »
« less
This content will become publicly available on December 7, 2026
An Evaluation of Interleaved Instruction Tuning on Semantic Reasoning Performance in an Audio MLLM
Standard training for Multi-modal Large Language Models (MLLMs) involves concatenating non-textual information, like vision or audio, with a text prompt. This approach may not encourage deep integration of modalities, limiting the model's ability to leverage the core language model's reasoning capabilities. This work examined the impact of interleaved instruction tuning in an audio MLLM, where audio tokens are interleaved within the prompt. Using the Listen, Think, and Understand (LTU) model as a testbed, we conduct an experiment using the Synonym and Hypernym Audio Reasoning Dataset (SHARD), our newly created reasoning benchmark for audio-based semantic reasoning focusing on synonym and hypernym recognition. Our findings show that while even zero-shot interleaved prompting improves performance on our reasoning tasks, a small amount of fine-tuning using interleaved training prompts improves the results further, however, at the expense of the MLLM's audio labeling ability.
more »
« less
- Award ID(s):
- 2228910
- PAR ID:
- 10657701
- Publisher / Repository:
- NeurIPS 2025 Multimodal Algorithmic Reasoning Workshop
- Date Published:
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Visual explanation (attention)-guided learning uses not only labels but also explanations to guide the model reasoning process. While visual attention-guided learning has shown promising results, it requires a large number of explanation annotations that are time-consuming to prepare. However, in many real-world situations, it is usually desired to prompt the model with visual attention without model retraining. For example, when doing AI-assisted cancer classification on a medical image, users (e.g., clinicians) can provide the AI model with visual attention prompts on which areas are indispensable and which are precluded. Despite its promising objectives, achieving visual attention-prompted prediction presents several major challenges: 1) How can the visual prompt be effectively integrated into the model's reasoning process? 2) How should the model handle samples that lack visual prompts? 3) What is the impact on the model's performance when a visual prompt is imperfect? This paper introduces a novel framework for visual attention prompted prediction and learning, utilizing visual prompts to steer the model's reasoning process. To improve performance in non-prompted situations and align it with prompted scenarios, we propose a co-training approach for both non-prompted and prompted models, ensuring they share similar parameters and activation. Additionally, for instances where the visual prompt does not encompass the entire input image, we have developed innovative attention prompt refinement methods. These methods interpolate the incomplete prompts while maintaining alignment with the model's explanations. Extensive experiments on four datasets demonstrate the effectiveness of our proposed framework in enhancing predictions for samples both with and without prompt.more » « less
-
Abstract In this paper, we evaluate the capability of transformer-based language models in making inferences over uncertain text that includes uncertain rules of reasoning. We cover both Pre-trained Language Models (PLMs) and generative Large Language Models (LLMs). Our evaluation results show that both generations of language models struggle with reasoning over uncertain text. We propose a novel end-to-end fine-tuning approach, Probabilistic Constraint Training (PCT), that utilizes probabilistic logical rules as constraints in the fine-tuning phase without relying on these rules in the inference stage. To assess the effectiveness of PCT, we utilize the related corpora and, additionally, create a new and more challenging benchmark that, unlike the previous ones, uses instance-specific rules. Our study demonstrates that PCT improves the transformer-based language model’s intrinsic reasoning and makes their probabilistic logical reasoning process more explicit and explainable. Furthermore, PCT equips these models to effectively handle novel situations, including higher reasoning depth, new domains, and complex probabilistic structures.more » « less
-
Existing approaches to automatic data transformation are insufficient to meet the requirements in many real-world scenarios, such as the building sector. First, there is no convenient interface for domain experts to provide domain knowledge easily. Second, they require significant training data collection overheads. Third, the accuracy suffers from complicated schema changes. To address these shortcomings, we present a novel approach that leverages the unique capabilities of large language models (LLMs) in coding, complex reasoning, and zero-shot learning to generate SQL code that transforms the source datasets into the target datasets. We demonstrate the viability of this approach by designing an LLM-based framework, termed SQLMorpher, which comprises a prompt generator that integrates the initial prompt with optional domain knowledge and historical patterns in external databases. It also implements an iterative prompt optimization mechanism that automatically improves the prompt based on flaw detection. The key contributions of this work include (1) pioneering an end-to-end LLM-based solution for data transformation, (2) developing a benchmark dataset of 105 real-world building energy data transformation problems, and (3) conducting an extensive empirical evaluation where our approach achieved 96% accuracy in all 105 problems. SQLMorpher demonstrates the effectiveness of utilizing LLMs in complex, domain-specific challenges, highlighting the potential of their potential to drive sustainable solutions.more » « less
-
Spatial sound reasoning is a fundamental human skill, enabling us to navigate and interpret our surroundings based on sound. In this paper we present BAT, which combines the spatial sound perception ability of a binaural acoustic scene analysis model with the natural language reasoning capabilities of a large language model (LLM) to replicate this innate ability. To address the lack of existing datasets of in-the-wild spatial sounds, we synthesized a binaural audio dataset using AudioSet and SoundSpaces 2.0. Next, we developed SpatialSoundQA, a spatial sound-based question-answering dataset, offering a range of QA tasks that train BAT in various aspects of spatial sound perception and reasoning. The acoustic front end encoder of BAT is a novel spatial audio encoder named Spatial Audio Spectrogram Transformer, or Spatial-AST, which by itself achieves strong performance across sound event detection, spatial localization, and distance estimation. By integrating Spatial-AST with LLaMA-2 7B model, BAT transcends standard Sound Event Localization and Detection (SELD) tasks, enabling the model to reason about the relationships between the sounds in its environment. Our experiments demonstrate BAT's superior performance on both spatial sound perception and reasoning, showcasing the immense potential of LLMs in navigating and interpreting complex spatial audio environments.more » « less
