Trust in Large Language Model chatbots depends not only on what these systems do but also on how their behavior is governed and communicated. We present Trust Mediator, a workbench that supports service owners in authoring and assessing principle sets for LLM-driven chatbots through persona-based exploration and structured scaffolds. To examine this workflow, we use three analytic lenses—specificity, coverage, and coherence—to characterize the principles produced. In an exploratory between-subjects study, we compared manual and assisted principle authoring. Participants in both conditions viewed principles as useful for governing and assessing chatbot behavior. Assisted authoring was generally perceived as more supportive and tended to broaden coverage. Manual authoring required more effort but yielded principles that were significantly more specific. These findings highlight complementary strengths of assisted and manual pathways, illustrating the value of treating principle sets as design objects within governance workflows. Beyond their analytic role in this study, the lenses also suggest opportunities for supporting the construction and inspection of principle sets.

Research on Large Language Models (LLMs) has recently witnessed an increasing interest in extending the models’ context size to better capture dependencies within long documents. While benchmarks have been proposed to assess long-range abilities, existing efforts primarily considered generic tasks that are not necessarily aligned with real-world applications. In contrast, we propose a new benchmark for long-context LLMs focused on a practical meeting assistant scenario in which the long contexts consist of transcripts obtained by automatic speech recognition, presenting unique challenges for LLMs due to the inherent noisiness and oral nature of such data. Our benchmark, ELITR-Bench, augments the existing ELITR corpus by adding 271 manually crafted questions with their ground-truth answers, as well as noisy versions of meeting transcripts altered to target different Word Error Rate levels. Our experiments with 12 long-context LLMs on ELITR-Bench confirm the progress made across successive generations of both proprietary and open models, and point out their discrepancies in terms of robustness to transcript noise. We also provide a thorough analysis of our GPT-4-based evaluation, including insights from a crowdsourcing study. Our findings indicate that while GPT-4’s scores align with human judges, its ability to distinguish beyond three score levels may be limited.

LLM-powered conversational assistants are often deployed in a one-size-fits-all manner, which fails to accommodate individual user preferences.Recently, LLM personalizationtailoring models to align with specific user preferences -has gained increasing attention as a way to bridge this gap.In this work, we specifically focus on a practical yet challenging setting where only a small set of preference annotations can be collected per user -a problem we define as Personalized Preference Alignment with Limited Data (PPALLI).To support research in this area, we introduce two datasets -DnD and ELIP -and benchmark a variety of alignment techniques on them.We further propose FaST, a highly parameter-efficient approach that leverages high-level features automatically discovered from the data, achieving the best overall performance.

As large language models (LLMs) are increasingly used across various applications, there is a growing need to control text generation to satisfy specific constraints or requirements. This raises a crucial question: Is it possible to guarantee strict constraint satisfaction in generated outputs while preserving the distribution of the original model as much as possible? We first define the ideal distribution — the one closest to the original model, which also always satisfies the expressed constraint — as the ultimate goal of guaranteed generation. We then state a fundamental limitation, namely that it is impossible to reach that goal through autoregressive training alone. This motivates the necessity of combining training-time and inference-time methods to enforce such guarantees. Based on this insight, we propose GUARD, a simple yet effective approach that combines an autoregressive proposal distribution with rejection sampling. Through GUARD’s theoretical properties, we show how controlling the KL divergence between a specific proposal and the target ideal distribution simultaneously optimizes inference speed and distributional closeness. To validate these theoretical concepts, we conduct extensive experiments on two text generation settings with hard-to-satisfy constraints: a lexical constraint scenario and a sentiment reversal scenario. These experiments show that GUARD achieves perfect constraint satisfaction while almost preserving the ideal distribution with highly improved inference efficiency. GUARD provides a principled approach to enforcing strict guarantees for LLMs without compromising their generative capabilities.

As language models (LMs) become more capable, it is increasingly important to align them with human preferences. However, the dominant paradigm for training Preference Models (PMs) for that purpose suffers from fundamental limitations, such as lack of transparency and scalability, along with susceptibility to overfitting the preference dataset. We propose Compositional Preference Models (CPMs), a novel PM framework that decomposes one global preference assessment into several interpretable features, obtains scalar scores for these features from a prompted LM, and aggregates these scores using a logistic regression classifier. Through these simple steps, CPMs allow to control which properties of the preference data are used to train the preference model and to build it based on features that are believed to underlie the human preference judgment. Our experiments show that CPMs not only improve generalization and are more robust to overoptimization than standard PMs, but also that best-of-n samples obtained using CPMs tend to be preferred over samples obtained using conventional PMs. Overall, our approach demonstrates the benefits of endowing PMs with priors about which features determine human preferences while relying on LM capabilities to extract those features in a scalable and robust way.

Aligning language models with preferences can be posed as approximating a target distribution representing some desired behavior. Existing approaches differ both in the functional form of the target distribution and the algorithm used to approximate it. For instance, Reinforcement Learning from Human Feedback (RLHF) corresponds to minimizing a reverse KL from an implicit target distribution arising from a KL penalty in the objective. On the other hand, Generative Distributional Control (GDC) has an explicit target distribution and minimizes a forward KL from it using the Distributional Policy Gradient (DPG) algorithm. In this paper, we propose a new approach, $f$-DPG, which allows the use of any $f$-divergence to approximate any target distribution that can be evaluated. $f$-DPG unifies both frameworks (RLHF, GDC) and the approximation methods (DPG, RL with KL penalties). We show the practical benefits of various choices of divergence objectives and demonstrate that there is no universally optimal objective but that different divergences present different alignment and diversity trade-offs. We show that Jensen-Shannon divergence strikes a good balance between these objectives, and frequently outperforms forward KL divergence by a wide margin, leading to significant improvements over prior work. These distinguishing characteristics between divergences persist as the model size increases, highlighting the importance of selecting appropriate divergence objectives.

We address the problem of unsupervised abstractive summarization of collections of user generated reviews through self-supervision and control.We propose a self-supervised setup that considers an individual document as a target summary for a set of similar documents.This setting makes training simpler than previous approaches by relying only on standard log-likelihood loss and mainstream models.We address the problem of hallucinations through the use of control codes, to steer the generation towards more coherent and relevant summaries.Our benchmarks on two English datasets against graph-based and recent neural abstractive unsupervised models show that our proposed method generates summaries with a superior quality and relevance, as well as a high sentiment and topic alignment with the input reviews.This is confirmed in our human evaluation which focuses explicitly on the faithfulness of generated summaries.We also provide an ablation study showing the importance of the control setup in controlling hallucinations.