An official website of the United States government
Abstract The phenomenon of ambiguity aversion suggests that people prefer options that offer precisely rather than imprecisely known chances of success. However, past work on people's responses to ambiguity in health treatment contexts found ambiguity seeking rather than aversion. The present work addressed whether such findings reflected a broad tendency for ambiguity seeking in health treatment contexts or whether specific attributions for ambiguity play a substantial role. In three studies, people choose between two treatment options that involved similar underlying probabilities, except that the probabilities for one option involved ambiguity. The attributions offered for the ambiguity played an important role in the results. For example, when the range of probabilities associated with an ambiguous treatment was attributed to the fact that different studies yield different results, participants tended to show ambiguity aversion or indifference. However, when the range was attributed to something that participants could control (e.g., regular application of a cream) or something about which they were overoptimistic (e.g., their immune system function), participants tended to show ambiguity seeking. Health professionals should be mindful of how people will interpret and use information about ambiguity when choosing among treatments.
more »
« less
This content will become publicly available on September 8, 2026
Can LLMs Disambiguate Grounded Language? The Case of PP Attachment
We study resolution of ambiguity in prepositional phrase attachment by Large Language Models in the zero-shot setting. We evaluate a strong “plausibility” baseline derived from token probabilities of descriptions encoding alternative attachments, and explore possible improvements using additional token probabilities that reflect aspects of information structure. Error analysis suggests directions for more sophisticated tools, common-sense reasoning, world knowledge, and additional context to better resolve ambiguity.
more »
« less
- PAR ID:
- 10639155
- Publisher / Repository:
- ACL anthology
- Date Published:
- Format(s):
- Medium: X
- Location:
- Varna, Bulgaria
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Language models produce a distribution over the next token; can we use this information to recover the prompt tokens? We consider the problem of language model inversion and show that next-token probabilities contain a surprising amount of information about the preceding text. Often we can recover the text in cases where it is hidden from the user, motivating a method for recovering unknown prompts given only the model's current distribution output. We consider a variety of model access scenarios, and show how even without predictions for every token in the vocabulary we can recover the probability vector through search. On Llama-2 7b, our inversion method reconstructs prompts with a BLEU of 59 and token-level F1 of 78 and recovers 27% of prompts exactly. Code for reproducing all experiments is available at this http URL.more » « less
-
As large language models (LLMs) become increasingly integrated into real-world applications such as code generation and chatbot assistance, extensive efforts have been made to align LLM behavior with human values, including safety. Jailbreak attacks, aiming to provoke unintended and unsafe behaviors from LLMs, remain a significant LLM safety threat. In this paper, we aim to defend LLMs against jailbreak attacks by introducing SafeDecoding, a safety-aware decoding strategy for LLMs to generate helpful and harmless responses to user queries. Our insight in developing SafeDecoding is based on the observation that, even though probabilities of tokens representing harmful contents outweigh those representing harmless responses, safety disclaimers still appear among the top tokens after sorting tokens by probability in descending order. This allows us to mitigate jailbreak attacks by identifying safety disclaimers and amplifying their token probabilities, while simultaneously attenuating the probabilities of token sequences that are aligned with the objectives of jailbreak attacks. We perform extensive experiments on five LLMs using six state-of-the-art jailbreak attacks and four benchmark datasets. Our results show that SafeDecoding significantly reduces attack success rate and harmfulness of jailbreak attacks without compromising the helpfulness of responses to benign user queries while outperforming six defense methods. Our code is publicly available at: https://github.com/uw-nsl/SafeDecodingmore » « less
-
Robust MDPs (RMDPs) can be used to compute policies with provable worst-case guarantees in reinforcement learning. The quality and robustness of an RMDP solution are determined by the ambiguity set—the set of plausible transition probabilities—which is usually constructed as a multi-dimensional confidence region. Existing methods construct ambiguity sets as confidence regions using concentration inequalities which leads to overly conservative solutions. This paper proposes a new paradigm that can achieve better solutions with the same robustness guarantees without using confidence regions as ambiguity sets. To incorporate prior knowledge, our algorithms optimize the size and position of ambiguity sets using Bayesian inference. Our theoretical analysis shows the safety of the proposed method, and the empirical results demonstrate its practical promisemore » « less
-
Large language model (LLM) decoding involves generating a sequence of tokens based on a given context, where each token is predicted one at a time using the model's learned probabilities. The typical autoregressive decoding method requires a separate forward pass through the model for each token generated, which is computationally inefficient and poses challenges for deploying LLMs in latency-sensitive scenarios.The main limitations of current decoding methods stem from their inefficiencies and resource demands. Existing approaches either necessitate fine-tuning smaller models, which is resource-intensive, or relying on fixed retrieval schemes to construct drafts for the next tokens, which lack adaptability and fail to generalize across different models and contexts.To address these issues, we introduce a novel methodology called Adaptix, which accelerates LLM decoding without requiring fine-tuning. Our approach involves an adaptive draft-verification process that evolves over time to improve efficiency. We utilize a tri-gram matrix-based LLM representation to dynamically approximate the output distribution of the LLM, allowing the model to adjust to changing token probabilities during the decoding process. Additionally, we implement a draft construction mechanism that effectively balances exploration and exploitation, ensuring that the drafts generated are both diverse and close to the true output distribution of the LLM.The importance of this design lies in its ability to optimize the draft distribution adaptively, leading to faster and more accurate decoding. Through extensive experiments on various benchmark datasets and LLM architectures, we demonstrate that Adaptix significantly accelerates the decoding process while maintaining high accuracy, making it suitable for deployment in a wide range of practical applications.more » « less
