The proliferation of open-source Large Language Models (LLMs) underscores the pressing need for evaluation methods. Existing works primarily rely on external evaluators, focusing on training and prompting strategies. However, a crucial aspect – model-aware glass-box features – is overlooked. In this study, we explore the utility of glass-box features under the scenario of self-evaluation, namely applying an LLM to evaluate its own output. We investigate various glass-box feature groups and discovered that the softmax distribution serves as a reliable quality indicator for self-evaluation. Experimental results on public benchmarks validate the feasibility of self-evaluation of LLMs using glass-box features.

Chain-of-Thought (CoT) has become a vital technique for enhancing the performance of Large Language Models (LLMs), attracting increasing attention from researchers. One stream of approaches focuses on the iterative enhancement of LLMs by continuously verifying and refining their reasoning outputs for desired quality. Despite its impressive results, this paradigm faces two critical issues: (1) Simple verification methods: The current paradigm relies solely on a single verification method. (2) Wrong Information Ignorance: Traditional paradigms directly ignore wrong information during reasoning and refine the logic paths from scratch each time. To address these challenges, we propose Wrong-of-Thought (WoT), which includes two core modules: (1) Multi-Perspective Verification: A multi-perspective verification method for accurately refining the reasoning process and result, and (2) Wrong Information Utilization: Utilizing wrong information to alert LLMs and reduce the probability of LLMs making same mistakes. Experiments on 8 popular datasets and 5 LLMs demonstrate that WoT surpasses all previous baselines. In addition, WoT exhibits powerful capabilities in difficult computation tasks.

The success of deep learning models on multi-hop fact verification has prompted researchers to understand the behavior behind their veracity. One feasible way is erasure search: obtaining the rationale by entirely removing a subset of input without compromising verification accuracy. Despite extensive exploration, current rationalization methods struggle to discern nuanced composition within the correlated evidence, which inevitably leads to noise rationalization in multi-hop scenarios. To address this issue, this paper explores the multi-granular rationale extraction method, aiming to realize the denoising rationalization for multi-hop fact verification. Specifically, given a pretrained veracity prediction model, two independent external explainers are introduced and trained collaboratively to enhance the discriminating ability by imposing varied constraints. Meanwhile, three key properties (Fidelity, Consistency, Salience) are introduced to regularize the denoising and faithful rationalization process. Additionally, a new Noiselessness metric is proposed to measure the purity of the rationales. Experimental results on three multi-hop fact verification datasets show that the proposed approach outperforms 12 baselines.

Medical entity disambiguation (MED) aims to ground medical mentions in text with ontological entities in knowledge bases (KBs). A notable challenge of MED is the long medical text usually contains multiple entities’ mentions with intricate correlations. However, limited by computation overhead, many existing methods consider only a single candidate entity mention during the disambiguation process. As such, they focus only on local MED optimal while ignoring the sole-mention disambiguation possibly boosted by richer context from other mentions’ disambiguating processes – missing global optimal on entity combination in the text. Motivated by this, we propose a new approach called Extractive Medical Entity Disambiguation with Memory Mechanism and Memorized Entity Information (M3E). Specifically, we reformulate MED as a text extraction task, which simultaneously accepts the context of medical mentions, all possible candidate entities, and entity definitions, and it is then trained to extract the text span corresponding to the correct entity. Upon our new formulation, 1) to alleviate the computation overhead from the enriched context, we devise a memory mechanism module that performs memory caching, retrieval, fusion and cross-network residual; and 2) to utilize the disambiguation clues from other mentions, we design an auxiliary disambiguation module that employs a gating mechanism to assist the disambiguation of remaining mentions. Extensive experiments on two benchmark datasets demonstrate the superiority of M3E over the state-of-the-art MED methods on all metrics.

With the growing complexity of fact verification tasks, the concern with “thoughtful” reasoning capabilities is increasing. However, recent fact verification benchmarks mainly focus on checking a narrow scope of semantic factoids within claims and lack an explicit logical reasoning process. In this paper, we introduce CHECKWHY, a challenging dataset tailored to a novel causal fact verification task: checking the truthfulness of the causal relation within claims through rigorous reasoning steps. CHECKWHY consists of over 19K “why” claim-evidence- argument structure triplets with supports, refutes, and not enough info labels. Each argument structure is composed of connected evidence, representing the reasoning process that begins with foundational evidence and progresses toward claim establishment. Through extensive experiments on state-of-the-art models, we validate the importance of incorporating the argument structure for causal fact verification. Moreover, the automated and human evaluation of argument structure generation reveals the difficulty in producing satisfying argument structure by fine-tuned models or Chain-of-Thought prompted LLMs, leaving considerable room for future improvements.

Medical entity disambiguation (MED) plays a crucial role in natural language processing and biomedical domains, which is the task of mapping ambiguous medical mentions to structured candidate medical entities from knowledge bases (KBs). However, existing methods for MED often fail to fully utilize the knowledge within medical KBs and overlook essential interactions between medical mentions and candidate entities, resulting in knowledge- and interaction-inefficient modeling and suboptimal disambiguation performance. To address these limitations, this paper proposes a novel approach, MED with Medical Mention Relation and Fine-grained Entity Knowledge (MMR-FEK). Specifically, MMR-FEK incorporates a mention relation fusion module and an entity knowledge fusion module, followed by an interaction module. The former employs a relation graph convolutional network to fuse mention relation information between medical mentions to enhance mention representations, while the latter leverages an attention mechanism to fuse synonym and type information of candidate entities to enhance entity representations. Afterwards, an interaction module is designed to employ a bidirectional attention mechanism to capture interactions between mentions and entities to generate the matching representation. Extensive experiments on two publicly available real-world datasets demonstrate MMR-FEK’s superiority over state-of-the-art(SOTA) MED baselines across all metrics. Our source code is publicly available.

Word sense disambiguation (WSD), identifying the most suitable meaning of ambiguous words in the given contexts according to a predefined sense inventory, is one of the most classical and challenging tasks in natural language processing. Benefiting from the powerful ability of deep neural networks, WSD has achieved a great advancement in recent years. Reformulating WSD as a text span extraction task is an effective approach, which accepts a sentence context of an ambiguous word together with all definitions of its candidate senses simultaneously, and requires to extract the text span corresponding with the right sense. However, the approach merely depends on a short definition to learn sense representation, which neglects abundant semantic knowledge from related senses and leads to data-inefficient learning and suboptimal WSD performance. To address the limitations, we propose a novel WSD method with Knowledge-Enhanced and Local Self-Attention-based Extractive Sense Comprehension (KELESC). Specifically, a knowledge-enhanced method is proposed to enrich semantic representation by incorporating additional examples and definitions of the related senses in WordNet. Then, in order to avoid the huge computing complexity induced by the additional information, a local self-attention mechanism is utilized to constrain attention to be local, which allows longer input texts without large-scale computing burdens. Extensive experimental results demonstrate that KELESC achieves better performance than baseline models on public benchmark datasets.

Sentence intention matching is vital for natural language understanding. Especially for Chinese sentence intention matching task, due to the ambiguity of Chinese words, semantic missing or semantic confusion are more likely to occur in the encoding process. Although the existing methods have enriched text representation through pre-trained word embedding to solve this problem, due to the particularity of Chinese text, different granularities of pre-trained word embedding will affect the semantic description of a piece of text. In this paper, we propose an effective approach that combines character-granularity and word-granularity features to perform sentence intention matching, and we utilize soft alignment attention to enhance the local information of sentences on the corresponding levels. The proposed method can capture sentence feature information from multiple perspectives and correlation information between different levels of sentences. By evaluating on BQ and LCQMC datasets, our model has achieved remarkable results, and demonstrates better or comparable performance with BERT-based models.

This paper reports the details of our submissions in the task 1 of SemEval 2017. This task aims at assessing the semantic textual similarity of two sentences or texts. We submit three unsupervised systems based on word embeddings. The differences between these runs are the various preprocessing on evaluation data. The best performance of these systems on the evaluation of Pearson correlation is 0.6887. Unsurprisingly, results of our runs demonstrate that data preprocessing, such as tokenization, lemmatization, extraction of content words and removing stop words, is helpful and plays a significant role in improving the performance of models.

This paper shows the details of our system submissions in the task 2 of SemEval 2017. We take part in the subtask 1 of this task, which is an English monolingual subtask. This task is designed to evaluate the semantic word similarity of two linguistic items. The results of runs are assessed by standard Pearson and Spearman correlation, contrast with official gold standard set. The best performance of our runs is 0.781 (Final). The techniques of our runs mainly make use of the word embeddings and the knowledge-based method. The results demonstrate that the combined method is effective for the computation of word similarity, while the word embeddings and the knowledge-based technique, respectively, needs more deeply improvement in details.