Feature attributions attempt to highlight what inputs drive predictive power. Good attributions or explanations are thus those that produce inputs that retain this predictive power; accordingly, evaluations of explanations score their quality of prediction. However, evaluations produce scores better than what appears possible from the values in the explanation for a class of explanations, called encoding explanations. Probing for encoding remains a challenge because there is no general characterization of what gives the extra predictive power. We develop a definition of encoding that identifies this extra predictive power via conditional dependence and show that the definition fits existing examples of encoding. This definition implies, in contrast to encoding explanations, that non-encoding explanations contain all the informative inputs used to produce the explanation, giving them a "what you see is what you get" property, which makes them transparent and simple to use. Next, we prove that existing scores (ROAR, FRESH, EVAL-X) do not rank non-encoding explanations above encoding ones, and develop STRIPE-X which ranks them correctly. After empirically demonstrating the theoretical insights, we use STRIPE-X to show that despite prompting an LLM to produce non-encoding explanations for a sentiment analysis task, the LLM-generated explanations encode.

Recent approaches to vision-language tasks are built on the remarkable capabilities of large vision-language models (VLMs). These models excel in zero-shot and few-shot learning, enabling them to learn new tasks without parameter updates. However, their primary challenge lies in their design, which primarily accommodates 2D input, thus limiting their effectiveness for medical images, particularly radiological images like MRI and CT, which are typically 3D. To bridge the gap between state-of-the-art 2D VLMs and 3D medical image data, we developed an innovative, one-pass, unsupervised representative slice selection method called Vote-MI, which selects representative 2D slices from 3D medical imaging. To evaluate the effectiveness of vote-MI when implemented with VLMs, we introduce BrainMD, a robust, multimodal dataset comprising 2,453 annotated 3D MRI brain scans with corresponding textual radiology reports and electronic health records.

The remarkable instruction-following capability of large language models (LLMs) has sparked a growing interest in automatically finding good prompts, i.e., prompt optimization. Most existing works follow the scheme of selecting from a pre-generated pool of candidate prompts. However, these designs mainly focus on the generation strategy, while limited attention has been paid to the selection method. Especially, the cost incurred during the selection (e.g., accessing LLM and evaluating the responses) is rarely explicitly considered. To overcome this limitation, this work provides a principled framework, TRIPLE, to efficiently perform prompt selection under an explicit budget constraint. TRIPLE is built on a novel connection established between prompt optimization and fixed-budget best arm identification (BAI-FB) in multi-armed bandits (MAB); thus, it is capable of leveraging the rich toolbox from BAI-FB systematically and also incorporating unique characteristics of prompt optimization. Extensive experiments on multiple well-adopted tasks using various LLMs demonstrate the remarkable performance improvement of TRIPLE over baselines while satisfying the limited budget constraints. As an extension, variants of TRIPLE are proposed to efficiently select examples for few-shot prompts, also achieving superior empirical performance.

The pre-trained Large Language Models (LLMs) can be adapted for many downstream tasks and tailored to align with human preferences through fine-tuning. Recent studies have discovered that LLMs can achieve desirable performance with only a small amount of high-quality data, suggesting that a large portion of the data in these extensive datasets is redundant or even harmful. Identifying high-quality data from vast datasets to curate small yet effective datasets has emerged as a critical challenge. In this paper, we introduce SHED, an automated dataset refinement framework based on Shapley value for instruction fine-tuning. SHED eliminates the need for human intervention or the use of commercial LLMs. Moreover, the datasets curated through SHED exhibit transferability, indicating they can be reused across different LLMs with consistently high performance. We conduct extensive experiments to evaluate the datasets curated by SHED. The results demonstrate SHED's superiority over state-of-the-art methods across various tasks and LLMs; notably, datasets comprising only 10% of the original data selected by SHED achieve performance comparable to or surpassing that of the full datasets.

Large Language Models (LLMs) are being used for a wide variety of tasks. While they are capable of generating human-like responses, they can also produce undesirable output including potentially harmful information, racist or sexist language, and hallucinations. Alignment methods are designed to reduce such undesirable output, via techniques such as fine-tuning, prompt engineering, and representation engineering. However, existing methods face several challenges: some require costly fine-tuning for every alignment task; some do not adequately remove undesirable concepts, failing alignment; some remove benign concepts, lowering the linguistic capabilities of LLMs. To address these issues, we propose Parsimonious Concept Engineering (PaCE), a novel activation engineering framework for alignment. First, to sufficiently model the concepts, we construct a large-scale concept dictionary in the activation space, in which each atom corresponds to a semantic concept. Given any alignment task, we instruct a concept partitioner to efficiently annotate the concepts as benign or undesirable. Then, at inference time, we decompose the LLM activations along the concept dictionary via sparse coding, to accurately represent the activations as linear combinations of benign and undesirable components.

With the rapid advancement of Natural Language Processing in recent years, numerous studies have shown that generic summaries generated by Large Language Models (LLMs) can sometimes surpass those annotated by experts, such as journalists, according to human evaluations. However, there is limited research on whether these generic summaries meet the individual needs of ordinary people. The biggest obstacle is the lack of human-annotated datasets from the general public. Existing work on personalized summarization often relies on pseudo datasets created from generic summarization datasets or controllable tasks that focus on specific named entities or other aspects, such as the length and specificity of generated summaries, collected from hypothetical tasks without the annotators' initiative. To bridge this gap, we propose a high-quality, personalized, manually annotated summarization dataset called PersonalSum. This dataset is the first to investigate whether the focus of public readers differs from the generic summaries generated by LLMs. It includes user profiles, personalized summaries accompanied by source sentences from given articles, and machine-generated generic summaries along with their sources. We investigate several personal signals -- entities/topics, plot, and structure of articles--that may affect the generation of personalized summaries using LLMs in a few-shot in-context learning scenario. Our preliminary results and analysis indicate that entities/topics are merely one of the key factors that impact the diverse preferences of users, and personalized summarization remains a significant challenge for existing LLMs.

To mitigate these risks, current evaluation benchmarks predominantly employ expert-designed contextual scenarios to assess how well LLMs align with human values. However, the labor-intensive nature of these benchmarks limits their test scope, hindering their ability to generalize to the extensive variety of open-world use cases and identify rare but crucial long-tail risks. Additionally, these static tests fail to adapt to the rapid evolution of LLMs, making it hard to evaluate timely alignment issues. To address these challenges, we propose ALI-Agent, an evaluation framework that leverages the autonomous abilities of LLM-powered agents to conduct in-depth and adaptive alignment assessments. ALI-Agent operates through two principal stages: Emulation and Refinement.

Despite the advanced intelligence abilities of large language models (LLMs) in various applications, they still face significant computational and storage demands. Knowledge Distillation (KD) has emerged as an effective strategy to improve the performance of a smaller LLM (i.e., the student model) by transferring knowledge from a high-performing LLM (i.e., the teacher model). Prevailing techniques in LLM distillation typically use a black-box model API to generate high-quality pretrained and aligned datasets, or utilize white-box distillation by altering the loss function to better transfer knowledge from the teacher LLM. However, these methods ignore the knowledge differences between the student and teacher LLMs across domains.

Machine unlearning is a promising solution for efficiently removing specific knowledge by post hoc modifying models. In this paper, we propose a Real-World Knowledge Unlearning benchmark (RWKU) for LLM unlearning. RWKU is designed based on the following three key factors: (1) For the task setting, we consider a more practical and challenging unlearning setting, where neither the forget corpus nor the retain corpus is accessible.

Evaluating large language models (LLMs) is challenging. Traditional ground-truth-based benchmarks fail to capture the comprehensiveness and nuance of real-world queries, while LLM-as-judge benchmarks suffer from grading biases and limited query quantity. Both of them may also become contaminated over time. User-facing evaluation, such as Chatbot Arena, provides reliable signals but is costly and slow. In this work, we propose MixEval, a new paradigm for establishing efficient, gold-standard LLM evaluation by strategically mixing off-the-shelf benchmarks.