VeriLA: A Human-Centered Evaluation Framework for Interpretable Verification of LLM Agent Failures

AI practitioners increasingly use large language model (LLM) agents in compound AI systems to solve complex reasoning tasks, these agent executions often fail to meet human standards, leading to errors that compromise the system's overall performance. Addressing these failures through human intervention is challenging due to the agents' opaque reasoning processes, misalignment with human expectations, the complexity of agent dependencies, and the high cost of manual inspection. This paper thus introduces a human-centered evaluation framework for Verifying LLM Agent failures (VeriLA), which systematically assesses agent failures to reduce human effort and make these agent failures interpretable to humans. The framework first defines clear expectations of each agent by curating human-designed agent criteria. Then, it develops a human-aligned agent verifier module, trained with human gold standards, to assess each agent's execution output. This approach enables granular evaluation of each agent's performance by revealing failures from a human standard, offering clear guidelines for revision, and reducing human cognitive load. Our case study results show that VeriLA is both interpretable and efficient in helping practitioners interact more effectively with the system. By upholding accountability in human-agent collaboration, VeriLA paves the way for more trustworthy and human-aligned compound AI systems.

Generative Uncertainty in Diffusion Models

Diffusion models have recently driven significant breakthroughs in generative modeling. While state-of-the-art models produce high-quality samples on average, individual samples can still be low quality. Detecting such samples without human inspection remains a challenging task. To address this, we propose a Bayesian framework for estimating generative uncertainty of synthetic samples. We outline how to make Bayesian inference practical for large, modern generative models and introduce a new semantic likelihood (evaluated in the latent space of a feature extractor) to address the challenges posed by high-dimensional sample spaces. Through our experiments, we demonstrate that the proposed generative uncertainty effectively identifies poor-quality samples and significantly outperforms existing uncertainty-based methods. Notably, our Bayesian framework can be applied post-hoc to any pretrained diffusion or flow matching model (via the Laplace approximation), and we propose simple yet effective techniques to minimize its computational overhead during sampling.

A Stronger Mixture of Low-Rank Experts for Fine-Tuning Foundation Models

In order to streamline the fine-tuning of foundation models, Low-Rank Adapters (LoRAs) have been substantially adopted across various fields, including instruction tuning and domain adaptation. The underlying concept of LoRA involves decomposing a full-rank matrix into the product of two lower-rank matrices, which reduces storage consumption and accelerates the training process. Furthermore, to address the limited expressive capacity of LoRA, the Mixture-of-Expert (MoE) has been introduced for incorporating multiple LoRA adapters. The integration of LoRA experts leads to a visible improvement across several downstream scenes. However, the mixture of LoRAs (MoE-LoRA) still exhibits its low robustness during tuning and inferring. Inspired by the Riemannian Preconditioners which train LoRA as a sub-space projector, we propose a new training strategy for MoE-LoRA, to stabilize and boost its feature learning procedure by multi-space projections. Examinations on SGD and AdamW optimizers demonstrate the effectiveness of our methodology. Source code is available at https://github.com/THUDM/MoELoRA_Riemannian.

Bridging the Gap: Transforming Natural Language Questions into SQL Queries via Abstract Query Pattern and Contextual Schema Markup

Large language models have demonstrated excellent performance in many tasks, including Text-to-SQL, due to their powerful in-context learning capabilities. They are becoming the mainstream approach for Text-to-SQL. However, these methods still have a significant gap compared to human performance, especially on complex questions. As the complexity of questions increases, the gap between questions and SQLs increases. We identify two important gaps: the structural mapping gap and the lexical mapping gap. To tackle these two gaps, we propose PAS-SQL, an efficient SQL generation pipeline based on LLMs, which alleviates gaps through Abstract Query Pattern (AQP) and Contextual Schema Markup (CSM). AQP aims to obtain the structural pattern of the question by removing database-related information, which enables us to find structurally similar demonstrations. CSM aims to associate database-related text span in the question with specific tables or columns in the database, which alleviates the lexical mapping gap. Experimental results on the Spider and BIRD datasets demonstrate the effectiveness of our proposed method. Specifically, PAS-SQL + GPT-4o sets a new state-of-the-art on the Spider benchmark with an execution accuracy of 87.9\%, and achieves leading results on the BIRD dataset with an execution accuracy of 64.67\%.

DataSciBench: An LLM Agent Benchmark for Data Science

This paper presents DataSciBench, a comprehensive benchmark for evaluating Large Language Model (LLM) capabilities in data science. Recent related benchmarks have primarily focused on single tasks, easily obtainable ground truth, and straightforward evaluation metrics, which limits the scope of tasks that can be evaluated. In contrast, DataSciBench is constructed based on a more comprehensive and curated collection of natural and challenging prompts for uncertain ground truth and evaluation metrics. We develop a semi-automated pipeline for generating ground truth (GT) and validating evaluation metrics. This pipeline utilizes and implements an LLM-based self-consistency and human verification strategy to produce accurate GT by leveraging collected prompts, predefined task types, and aggregate functions (metrics). Furthermore, we propose an innovative Task - Function - Code (TFC) framework to assess each code execution outcome based on precisely defined metrics and programmatic rules. Our experimental framework involves testing 6 API-based models, 8 open-source general models, and 9 open-source code generation models using the diverse set of prompts we have gathered. This approach aims to provide a more comprehensive and rigorous evaluation of LLMs in data science, revealing their strengths and weaknesses. Experimental results demonstrate that API-based models outperform open-sourced models on all metrics and Deepseek-Coder-33B-Instruct achieves the highest score among open-sourced models. We release all code and data at https://github.com/THUDM/DataSciBench.

Parameter-Efficient Fine-Tuning for Foundation Models

This survey delves into the realm of Parameter-Efficient Fine-Tuning (PEFT) within the context of Foundation Models (FMs). PEFT, a cost-effective fine-tuning technique, minimizes parameters and computational complexity while striving for optimal downstream task performance. FMs, like ChatGPT, DALL-E, and LLaVA specialize in language understanding, generative tasks, and multimodal tasks, trained on diverse datasets spanning text, images, and videos. The diversity of FMs guides various adaptation strategies for PEFT. Therefore, this survey aims to provide a comprehensive overview of PEFT techniques applied to diverse FMs and address critical gaps in understanding the techniques, trends, and applications. We start by providing a detailed development of FMs and PEFT. Subsequently, we systematically review the key categories and core mechanisms of PEFT across diverse FMs to offer a comprehensive understanding of trends. We also explore the most recent applications across various FMs to demonstrate the versatility of PEFT, shedding light on the integration of systematic PEFT methods with a range of FMs. Furthermore, we identify potential research and development directions for improving PEFTs in the future. This survey provides a valuable resource for both newcomers and experts seeking to understand and use the power of PEFT across FMs. All reviewed papers are listed at \url{https://github.com/THUDM/Awesome-Parameter-Efficient-Fine-Tuning-for-Foundation-Models}.

ZeroFlow: Overcoming Catastrophic Forgetting is Easier than You Think

Backpropagation provides a generalized configuration for overcoming catastrophic forgetting. Like, SGD and Adam are commonly used for weight updates in continual learning and continual pre-training. In practice, permission to access gradient information is not always granted (the gradient ban), such as black-box APIs, hardware limitations, and non-differentiable systems. To bridge this gap, we introduce the first benchmark ZeroFlow to evaluate gradient-free optimization algorithms for overcoming forgetting. This benchmark examines a suite of forward pass methods across multiple methods, forgetting scenarios, and datasets. We find that forward passes alone are enough to overcome forgetting. Our findings reveal new optimization principles that highlight the potential of forward-pass in mitigating forgetting, managing task conflicts, and reducing memory demands, alongside novel enhancements that further mitigate forgetting with just one forward pass. This work provides essential insights and tools for advancing forward pass methods to overcome forgetting.

SPaR: Self-Play with Tree-Search Refinement to Improve Instruction-Following in Large Language Models

Instruction-following is a fundamental capability of language models, requiring the model to recognize even the most subtle requirements in the instructions and accurately reflect them in its output. Such an ability is well-suited for and often optimized by preference learning. However, existing methods often directly sample multiple independent responses from the model when creating preference pairs. Such practice can introduce content variations irrelevant to whether the instruction is precisely followed (e.g., different expressions about the same semantic), interfering with the goal of teaching models to recognize the key differences that lead to improved instruction following. In light of this, we introduce SPaR, a self-play framework integrating tree-search self-refinement to yield valid and comparable preference pairs free from distractions. By playing against itself, an LLM employs a tree-search strategy to refine its previous responses with respect to the instruction while minimizing unnecessary variations. Our experiments show that a LLaMA3-8B model, trained over three iterations guided by SPaR, surpasses GPT-4-Turbo on the IFEval benchmark without losing general capabilities. Furthermore, SPaR demonstrates promising scalability and transferability, greatly enhancing models like GLM-4-9B and LLaMA3-70B. We also identify how inference scaling in tree search would impact model performance. Our code and data are publicly available at https://github.com/thu-coai/SPaR.

Object-Focused Data Selection for Dense Prediction Tasks

Dense prediction tasks such as object detection and segmentation require high-quality labels at pixel level, which are costly to obtain. Recent advances in foundation models have enabled the generation of autolabels, which we find to be competitive but not yet sufficient to fully replace human annotations, especially for more complex datasets. Thus, we consider the challenge of selecting a representative subset of images for labeling from a large pool of unlabeled images under a constrained annotation budget. This task is further complicated by imbalanced class distributions, as rare classes are often underrepresented in selected subsets. We propose object-focused data selection (OFDS) which leverages object-level representations to ensure that the selected image subsets semantically cover the target classes, including rare ones. We validate OFDS on PASCAL VOC and Cityscapes for object detection and semantic segmentation tasks. Our experiments demonstrate that prior methods which employ image-level representations fail to consistently outperform random selection. In contrast, OFDS consistently achieves state-of-the-art performance with substantial improvements over all baselines in scenarios with imbalanced class distributions. Moreover, we demonstrate that pre-training with autolabels on the full datasets before fine-tuning on human-labeled subsets selected by OFDS further enhances the final performance.

Argumentative Experience: Reducing Confirmation Bias on Controversial Issues through LLM-Generated Multi-Persona Debates

Large language models (LLMs) are enabling designers to give life to exciting new user experiences for information access. In this work, we present a system that generates LLM personas to debate a topic of interest from different perspectives. How might information seekers use and benefit from such a system? Can centering information access around diverse viewpoints help to mitigate thorny challenges like confirmation bias in which information seekers over-trust search results matching existing beliefs? How do potential biases and hallucinations in LLMs play out alongside human users who are also fallible and possibly biased? Our study exposes participants to multiple viewpoints on controversial issues via a mixed-methods, within-subjects study. We use eye-tracking metrics to quantitatively assess cognitive engagement alongside qualitative feedback. Compared to a baseline search system, we see more creative interactions and diverse information-seeking with our multi-persona debate system, which more effectively reduces user confirmation bias and conviction toward their initial beliefs. Overall, our study contributes to the emerging design space of LLM-based information access systems, specifically investigating the potential of simulated personas to promote greater exposure to information diversity, emulate collective intelligence, and mitigate bias in information seeking.