As Large Language Models (LLMs) scale up and gain powerful Chain-of-Thoughts (CoTs) reasoning abilities, practical resource constraints drive efforts to distill these capabilities into more compact Smaller Language Models (SLMs). We find that CoTs consist mainly of simple reasoning forms, with a small proportion (4.7%) of key reasoning steps that truly impact conclusions. However, previous distillation methods typically involve supervised fine-tuning student SLMs only on correct CoTs data produced by teacher LLMs, resulting in students struggling to learn the key, instead imitating the teacher’s reasoning forms and making errors or omissions in reasoning. To address these issues, drawing an analogy to human learning, where analyzing mistakes according to correct solutions often reveals the crucial steps leading to successes or failures, we propose mistakE-Driven key reasonIng step distillaTion (EDIT), a novel method that further aids SLMs learning key reasoning steps rather than mere simple fine-tuning. Firstly, to expose the crucial steps in CoTs, we carefully design specific prompts to generate dual CoTs data with similar reasoning paths but divergent conclusions. Then, we apply the minimum edit distance algorithm on the dual CoTs data to locate these key steps and optimize the likelihood on these tokens. Extensive experiments and analysis validate the effectiveness of EDIT across both in-domain(IND) and out-of-domain(OOD) benchmark reasoning datasets.

Knowledge Graphs (KGs) can serve as reliable knowledge sources for question answering (QA) due to their structured representation of knowledge. Existing research on the utilization of KG for large language models (LLMs) prevalently relies on subgraph retriever or iterative prompting, overlooking the potential synergy of LLMs’ step-wise reasoning capabilities and KGs’ structural nature. In this paper, we present DoG (Decoding on Graph), a novel framework that facilitates a deep synergy between LLMs and KGs. We first define a concept, well-formed chain, which consists of a sequence of interrelated fact triplets on the KGs, starting from question entities and leading to answers. We argue that this concept can serve as a principle for making faithful and sound reasoning for KGQA. To enable LLMs to generate well-formed chains, we propose graph-aware constrained decoding, in which a constraint derived from the topology of the KG regulates the decoding process of the LLMs. This constrained decoding method ensures the generation of well-formed chains while making full use of the step-wise reasoning capabilities of LLMs. Based on the above, DoG, a training-free approach, is able to provide faithful and sound reasoning trajectories grounded on the KGs. Experiments across various KGQA tasks with different background KGs demonstrate that DoG achieves superior and robust performance. DoG also shows general applicability with various open-source LLMs.

Autonomous agents powered by large language models (LLMs) hold significant potential across various domains. The Reflection framework is designed to help agents learn from past mistakes in complex tasks. While previous research has shown that reflection can enhance performance, our investigation reveals a key limitation: meaningful self-reflection primarily occurs at the beginning of iterations, with subsequent attempts failing to produce further improvements. We term this phenomenon “Early Stop Reflection,” where the reflection process halts prematurely, limiting the agent’s ability to engage in continuous learning. To address this, we propose the DORA method (Dynamic and Optimized Reflection Advice), which generates task-adaptive and diverse reflection advice. DORA introduces an external open-source small language model (SLM) that dynamically generates prompts for the reflection LLM. The SLM uses feedback from the agent and optimizes the prompt generation process through a non-gradient Bayesian Optimization (BO) algorithm, ensuring the reflection process evolves and adapts over time. Our experiments in the MiniWoB++ and Alfworld environments confirm that DORA effectively mitigates the “Early Stop Reflection” issue, enabling agents to maintain iterative improvements and boost performance in long-term, complex tasks. Code are available at https://anonymous.4open.science/r/DORA-44FB/.

Improving context faithfulness in large language models is essential for developing trustworthy retrieval augmented generation systems and mitigating hallucinations, especially in long-form question answering (LFQA) tasks or scenarios involving knowledge conflicts. Existing methods either intervene LLMs only at inference without addressing their inherent limitations or overlook the potential for self-improvement. In this paper, we introduce GenDiE(Generate, Discriminate, Evolve), a novel self-evolving framework that enhances context faithfulness through fine-grained sentence-level optimization. GenDiE combines both generative and discriminative training, equipping LLMs with self-generation and self-scoring capabilities to facilitate iterative self-evolution. This supports both data construction for model alignment and score-guided search during inference. Furthermore, by treating each sentence in a response as an independent optimization unit, GenDiE effectively addresses the limitations of previous approaches that optimize at the holistic answer level, which may miss unfaithful details. Experiments on ASQA (in-domain LFQA) and ConFiQA (out-of-domain counterfactual QA) datasets demonstrate that GenDiE surpasses various baselines in both faithfulness and correctness, and exhibits robust performance for domain adaptation.

Query rewriting is a crucial technique for passage retrieval in open-domain conversational question answering (CQA). It decontexualizes conversational queries into self-contained questions suitable for off-the-shelf retrievers. Existing methods attempt to incorporate retriever’s preference during the training of rewriting models. However, these approaches typically rely on extensive annotations such as in-domain rewrites and/or relevant passage labels, limiting the models’ generalization and adaptation capabilities. In this paper, we introduce AdaQR (Adaptive Query Rewriting), a framework for training query rewriting models with limited rewrite annotations from seed datasets and completely no passage label. Our approach begins by fine-tuning compact large language models using only 10% of rewrite annotations from the seed dataset training split. The models are then utilized to self-sample rewrite candidates for each query instance, further eliminating the expense for human labeling or larger language model prompting often adopted in curating preference data. A novel approach is then proposed to assess retriever’s preference for these candidates with the probability of answers conditioned on the conversational query by marginalizing the Top-K passages. This serves as the reward for optimizing the rewriter further using Direct Preference Optimization (DPO), a process free of rewrite and retrieval annotations. Experimental results on four open-domain CQA datasets demonstrate that AdaQR not only enhances the in-domain capabilities of the rewriter with limited annotation requirement, but also adapts effectively to out-of-domain datasets.

Large language models (LLMs) exhibit enhanced reasoning at larger scales, driving efforts to distill these capabilities into smaller models via teacher-student learning.Previous works simply fine-tune student models on teachers’ generated Chain-of-Thoughts (CoTs) data. Although these methods enhance in-domain (IND) reasoning performance, they struggle to generalize to out-of-domain (OOD) tasks.We believe that the widespread spurious correlations between questions and answers may lead the model to preset a specific answer which restricts the diversity and generalizability of its reasoning process.In this paper, we propose Cascading Decomposed CoTs Distillation (CasCoD) to address these issues by decomposing the traditional single-step learning process into two cascaded learning steps. Specifically, by restructuring the training objectives—removing the answer from outputs and concatenating the question with the rationale as input—CasCoD’s two-step learning process ensures that students focus on learning rationales without interference from the preset answers, thus improving reasoning generalizability. Extensive experiments demonstrate the effectiveness of CasCoD on both IND and OOD benchmark reasoning datasets

Text-to-SQL aims to convert natural language into structured query language, which is a challenging task. Current research focuses mainly on read operations and ignores other aspects of database operations such as create, update, and delete operations. The benchmark datasets as well as models that have been proposed also fail to cover these operations, limiting the development and practical applications in the field. To bridge this gap, we propose CRUDSQL, a large-scale cross-domain single-table CRUD operations Chinese Text-to-SQL dataset. The dataset contains 10,000 question/SQL pairs involving 625 tables from different domains. To support further research on this dataset, we also propose a baseline method, CRUDParser, which employs a two-phase approach based on BERT and T5 for SQL generation and incorporates two strategies, value matching, and value prompting, for interacting with databases to further improve the performance. The experimental results show that the new operation types bring different challenges for future research, and our approach achieves 67.08% and 83.8% exact set matching accuracy under both read and delete operations in the test set, but only 49.6% and 61.8% under create and update operations. We believe that the proposal of CRUDSQL as well as CRUDParser can provide new directions and possibilities for research and practical applications in the field of Text-to-SQL. The dataset is published at https://github.com/bizard-lab/CRUDSQL.

Neural network models are vulnerable to adversarial examples, and adversarial transferability further increases the risk of adversarial attacks. Current methods based on transferability often rely on substitute models, which can be impractical and costly in real-world scenarios due to the unavailability of training data and the victim model’s structural details. In this paper, we propose a novel approach that directly constructs adversarial examples by extracting transferable features across various tasks. Our key insight is that adversarial transferability can extend across different tasks. Specifically, we train a sequence-to-sequence generative model named CT-GAT (Cross-Task Generative Adversarial Attack) using adversarial sample data collected from multiple tasks to acquire universal adversarial features and generate adversarial examples for different tasks.We conduct experiments on ten distinct datasets, and the results demonstrate that our method achieves superior attack performance with small cost.

We study model extraction attacks in natural language processing (NLP) where attackers aim to steal victim models by repeatedly querying the open Application Programming Interfaces (APIs). Recent works focus on limited-query budget settings and adopt random sampling or active learning-based sampling strategies on publicly available, unannotated data sources. However, these methods often result in selected queries that lack task relevance and data diversity, leading to limited success in achieving satisfactory results with low query costs. In this paper, we propose MeaeQ (Model extraction attack with efficient Queries), a straightforward yet effective method to address these issues. Specifically, we initially utilize a zero-shot sequence inference classifier, combined with API service information, to filter task-relevant data from a public text corpus instead of a problem domain-specific dataset. Furthermore, we employ a clustering-based data reduction technique to obtain representative data as queries for the attack. Extensive experiments conducted on four benchmark datasets demonstrate that MeaeQ achieves higher functional similarity to the victim model than baselines while requiring fewer queries.

Building and maintaining end-to-end task bots using minimal human effort is a long-standing challenge in dialog research. In this work, we introduce SGP-TOD, Schema-Guided Prompting for building Task-Oriented Dialog systems effortlessly based on large language models (LLMs). Utilizing the predefined task schema, i.e., belief instruction and dialog policy, we instruct fixed LLMs to generate appropriate responses on novel tasks, without the need for training data. Specifically, SGP-TOD comprises three components: an LLM for interacting with users, a Dialog State Tracking (DST) Prompter to aid the LLM in tracking dialog states with the given belief instruction, and a Policy Prompter to direct the LLM to generate proper responses adhering to the provided dialog policy. Experimental results on Multiwoz, RADDLE, and STAR datasets show that our training-free strategy, SGP-TOD, yields state-of-the-art (SOTA) zero-shot performance, significantly surpassing the few-shot approaches. In a domain-extension setting, SGP-TOD aptly adapts to new functionalities by merely adding supplementary schema rules. We make our code and data publicly available.

MultiDoc2Dial presents an important challenge on modeling dialogues grounded with multiple documents. This paper proposes a pipeline system of “retrieve, re-rank, and generate”, where each component is individually optimized. This enables the passage re-ranker and response generator to fully exploit training with ground-truth data. Furthermore, we use a deep cross-encoder trained with localized hard negative passages from the retriever. For the response generator, we use grounding span prediction as an auxiliary task to be jointly trained with the main task of response generation. We also adopt a passage dropout and regularization technique to improve response generation performance. Experimental results indicate that the system clearly surpasses the competitive baseline and our team CPII-NLP ranked 1st among the public submissions on ALL four leaderboards based on the sum of F1, SacreBLEU, METEOR and RougeL scores.

Aspect term extraction aims to extract aspect terms from review texts as opinion targets for sentiment analysis. One of the big challenges with this task is the lack of sufficient annotated data. While data augmentation is potentially an effective technique to address the above issue, it is uncontrollable as it may change aspect words and aspect labels unexpectedly. In this paper, we formulate the data augmentation as a conditional generation task: generating a new sentence while preserving the original opinion targets and labels. We propose a masked sequence-to-sequence method for conditional augmentation of aspect term extraction. Unlike existing augmentation approaches, ours is controllable and allows to generate more diversified sentences. Experimental results confirm that our method alleviates the data scarcity problem significantly. It also effectively boosts the performances of several current models for aspect term extraction.

One of the remaining challenges for aspect term extraction in sentiment analysis resides in the extraction of phrase-level aspect terms, which is non-trivial to determine the boundaries of such terms. In this paper, we aim to address this issue by incorporating the span annotations of constituents of a sentence to leverage the syntactic information in neural network models. To this end, we first construct a constituency lattice structure based on the constituents of a constituency tree. Then, we present two approaches to encoding the constituency lattice using BiLSTM-CRF and BERT as the base models, respectively. We experimented on two benchmark datasets to evaluate the two models, and the results confirm their superiority with respective 3.17 and 1.35 points gained in F1-Measure over the current state of the art. The improvements justify the effectiveness of the constituency lattice for aspect term extraction.