Abstract:Large language models (LLMs) have shown increasing capability in problem-solving and decision-making, largely based on the step-by-step chain-of-thought reasoning processes. However, it has been increasingly challenging to evaluate the reasoning capability of LLMs. Concretely, existing outcome-based benchmarks begin to saturate and become less sufficient to monitor the progress. To this end, we present a process-based benchmark MR-BEN that demands a meta reasoning skill, where LMs are asked to locate and analyse potential errors in automatically generated reasoning steps. MR-BEN is a comprehensive benchmark comprising 5,975 questions collected from human experts, covering various subjects such as physics, chemistry, logic, coding, and more. Through our designed metrics for assessing meta-reasoning on this benchmark, we identify interesting limitations and weaknesses of current LLMs (open-source and closed-source models). For example, open-source models are seemingly comparable to GPT-4 on outcome-based benchmarks, but they lag far behind on our benchmark, revealing the underlying reasoning capability gap between them. Our dataset and codes are available on https://randolph-zeng.github.io/Mr-Ben.github.io/.
Abstract:Large language models (LLMs) have shown promising abilities as cost-effective and reference-free evaluators for assessing language generation quality. In particular, pairwise LLM evaluators, which compare two generated texts and determine the preferred one, have been employed in a wide range of applications. However, LLMs exhibit preference biases and worrying sensitivity to prompt designs. In this work, we first reveal that the predictive preference of LLMs can be highly brittle and skewed, even with semantically equivalent instructions. We find that fairer predictive preferences from LLMs consistently lead to judgments that are better aligned with humans. Motivated by this phenomenon, we propose an automatic Zero-shot Evaluation-oriented Prompt Optimization framework, ZEPO, which aims to produce fairer preference decisions and improve the alignment of LLM evaluators with human judgments. To this end, we propose a zero-shot learning objective based on the preference decision fairness. ZEPO demonstrates substantial performance improvements over state-of-the-art LLM evaluators, without requiring labeled data, on representative meta-evaluation benchmarks. Our findings underscore the critical correlation between preference fairness and human alignment, positioning ZEPO as an efficient prompt optimizer for bridging the gap between LLM evaluators and human judgments.
Abstract:The success of deep learning in transient stability assessment (TSA) heavily relies on high-quality training data. However, the label information in TSA datasets is vulnerable to contamination through false label injection (FLI) cyberattacks, resulting in degraded performance of deep TSA models. To address this challenge, a Multi-Module Robust TSA method (MMR) is proposed to rectify the supervised training process misguided by FLI in an unsupervised manner. In MMR, a supervised classification module and an unsupervised clustering module are alternatively trained to improve the clustering friendliness of representation leaning, thereby achieving accurate clustering assignments. Leveraging the clustering assignments, we construct a training label corrector to rectify the injected false labels and progressively enhance robustness and resilience against FLI. However, there is still a gap on accuracy and convergence speed between MMR and FLI-free deep TSA models. To narrow this gap, we further propose a human-in-the-loop training strategy, named MMR-HIL. In MMR-HIL, potential false samples can be detected by modeling the training loss with a Gaussian distribution. From these samples, the most likely false samples and most ambiguous samples are re-labeled by a TSA experts guided bi-directional annotator and then subjected to penalized optimization, aimed at improving accuracy and convergence speed. Extensive experiments indicate that MMR and MMR-HIL both exhibit powerful robustness against FLI in TSA performance. Moreover, the contaminated labels can also be effectively corrected, demonstrating superior resilience of the proposed methods.
Abstract:Large Language Models (LLMs) have demonstrated promising capabilities as automatic evaluators in assessing the quality of generated natural language. However, LLMs still exhibit biases in evaluation and often struggle to generate coherent evaluations that align with human assessments. In this work, we first conduct a systematic study of the misalignment between LLM evaluators and human judgement, revealing that existing calibration methods aimed at mitigating biases are insufficient for effectively aligning LLM evaluators. Inspired by the use of preference data in RLHF, we formulate the evaluation as a ranking problem and introduce Pairwise-preference Search (PairS), an uncertainty-guided search method that employs LLMs to conduct pairwise comparisons and efficiently ranks candidate texts. PairS achieves state-of-the-art performance on representative evaluation tasks and demonstrates significant improvements over direct scoring. Furthermore, we provide insights into the role of pairwise preference in quantifying the transitivity of LLMs and demonstrate how PairS benefits from calibration.
Abstract:In light of recent advances in large language models (LLMs), the expectations for the next generation of virtual assistants include enhanced naturalness and adaptability across diverse usage scenarios. However, the creation of high-quality annotated data for Task-Oriented Dialog (TOD) is recognized to be slow and costly. To address these challenges, we introduce Task-Oriented Automatic Dialogs (TOAD), a novel and scalable TOD dataset along with its automatic generation pipeline. The TOAD dataset simulates realistic app context interaction and provide a variety of system response style options. Two aspects of system response styles are considered, verbosity level and users' expression mirroring. We benchmark TOAD on two response generation tasks and the results show that modelling more verbose or responses without user expression mirroring is more challenging.
Abstract:Recent large language models (LLMs) have shown remarkable performance in aligning generated text with user intentions across various tasks. When it comes to long-form text generation, there has been a growing interest in generation from a discourse coherence perspective. However, existing lexical or semantic metrics such as BLEU, ROUGE, BertScore cannot effectively capture the discourse coherence. The development of discourse-specific automatic evaluation methods for assessing the output of LLMs warrants greater focus and exploration. In this paper, we present a novel automatic metric designed to quantify the discourse divergence between two long-form articles. Extensive experiments on three datasets from representative domains demonstrate that our metric aligns more closely with human preferences and GPT-4 coherence evaluation, outperforming existing evaluation methods.
Abstract:Instruction-tuned large language models have shown remarkable performance in aligning generated text with user intentions across various tasks. However, maintaining human-like discourse structure in the generated text remains a challenging research question. In this paper, we propose Instruct-SCTG, a flexible and effective sequential framework that harnesses instruction-tuned language models to generate structurally coherent text in both fine-tuned and zero-shot setups. Our framework generates articles in a section-by-section manner, aligned with the desired human structure using natural language instructions. Furthermore, we introduce a new automatic metric that measures discourse divergence in a fuzzy manner. Extensive experiments on three datasets from representative domains of news and recipes demonstrate the state-of-the-art performance of our framework in imposing discourse structure during text generation, as verified by both automatic and human evaluation. Our code will be available on Github.
Abstract:Recent pre-trained language models have shown promising capabilities in generating fluent and realistic natural language text. However, generating multi-sentence text with global content planning has been a long-existing research question. Current approaches for controlled text generation can hardly address this issue, as they usually condition on single known control attributes. In this study, we propose a low-cost yet effective framework which explicitly models the global content plan of the generated text. Specifically, it optimizes the joint distribution of the natural language sequence and the global content plan in a plug-and-play manner. We conduct extensive experiments on the well-established Recipe1M+ benchmark. Both automatic and human evaluations verify that our model achieves the state-of-the-art performance on the task of recipe generation
Abstract:Functional Distributional Semantics is a recently proposed framework for learning distributional semantics that provides linguistic interpretability. It models the meaning of a word as a binary classifier rather than a numerical vector. In this work, we propose a method to train a Functional Distributional Semantics model with grounded visual data. We train it on the Visual Genome dataset, which is closer to the kind of data encountered in human language acquisition than a large text corpus. On four external evaluation datasets, our model outperforms previous work on learning semantics from Visual Genome.