Knowledge editing algorithms have emerged as an efficient paradigm for updating parametric facts in Large Language Models (LLMs) without full retraining. However, existing benchmarks primarily focus on isolated, static facts. While a few temporal datasets exist, they are frequently bottlenecked by coarse temporal granularity, severe relation imbalance, and a critical lack of a priori knowledge validation, which can conflate inherent model deficits with editing failures. To address these structural limitations, we introduce TRACE, a fine-grained Temporal Reasoning And Chronological Editing benchmark. Using exact-match a priori probing, TRACE dynamically constructs model-specific, non-mutually exclusive chronological event chains. Featuring over 16,000 customized test cases across two distinct operational setups (Modify and Insert), TRACE provides a comprehensive 7-dimensional metric matrix. It extends simple fact recall with relation-constrained counterfactual injection and sequential temporal logic probes that test timeline integrity. Our evaluation of 6 representative baselines (FT, IKE, ROME, GRACE, LoRA, WISE) on Llama-3.2-3B and GPT-J-6B identifies widespread temporal interference in the studied settings: editing a single timestamp often sharply degrades recall of adjacent historical and future facts. The consistently weak bidirectional inference after editing further suggests that the evaluated methods tend to introduce shallow, syntax-sensitive patches rather than integrate the edit into a chronological network. Our code, the EasyEdit snapshot used for the experiments, and the benchmark are publicly available.

Recent advances in large language models (LLMs) have shown strong potential for knowledge graph completion (KGC). However, existing LLM-based approaches often struggle to effectively capture the structural information in knowledge graphs (KGs), leading to suboptimal reasoning performance. To address this challenge, we propose a Multi-layer Aligned Knowledge Injection (MAKI) model, a novel method that tightly integrates structured KG information into LLMs through multi-layer alignment. Specifically, we first leverage LLMs to encode the textual information of entities and relations, obtaining their semantic representations across multiple hidden layers. We then introduce a multi-layer aligned structure learning module, which uses graph neural networks (GNNs) to learn relational structures while aligning with the corresponding LLM layers to bridge the gap between structural and semantic spaces. Finally, a gated fusion mechanism is used to inject the structured knowledge into the LLM for reasoning over candidate triples. Experimental results on various benchmark datasets demonstrate that the proposed MAKI outperforms existing state-of-the-art methods.

Large Audio Language Models (LALMs) are increasingly applied in affective and conversational AI, yet their ability to align with human empathetic judgments in spoken dialogue remains poorly understood. We introduce BAEmpEval, a multimodal evaluation framework for analyzing empathetic response judgment in instruction-following LALMs under subjective conversational settings. The framework contains 5,436 dialogue pairs collected from three speech sources, where models evaluate listener responses across six fine-grained dimensions spanning both vocal perception (Vocal Warmth, Perceived Genuineness, Prosodic Attunement) and conversational response strategies (Exploration, Interpretation, Emotional Reaction). Experiments reveal a challenging setting: experiments across recent LALMs reveal uneven alignment with human relational judgments across empathy dimensions and speech conditions. Our findings suggest that empathetic response judgment remains a challenging multidimensional perceptual task for current LALMs and highlight the importance of evaluation frameworks that account for subjective relational variability. We release the dataset and evaluation code to support future research in human-centric AI.

In the context of today’s high-pressure, aging society, the demand for large-scale emotional models capable of providing empathetic support is more critical than ever. However, existing benchmarks fail to simultaneously achieve ecological validity, signal clarity, and reliable fine-grained labeling. We introduce EmoS, a high-fidelity bilingual benchmark designed to resolve the limitations of ecological validity and noise in existing datasets by combining strictly filtered static slices with a dynamic Streaming Monologue subset. Supported by a rigorous dual-layer human annotation pipeline, EmoS provides trusted ground truth that captures continuous emotional evolution. Empirical results show that fine-tuning MLLMs (multimodal large language models) on EmoS yields significant gains over zero-shot baselines, laying the foundation for the training and evaluation of future emotion recognition models and empathy models. The dataset and code are publicly available at https://github.com/NLP2CT/EmoS.

Large Audio Language Models (LALMs) exhibit strong speech understanding ability, yet it remains unclear whether they meaningfully utilize subtle speech-conditioned emotional cues beyond coarse acted emotions. In this work, we investigate whether the evaluated LALMs exhibit measurable sensitivity to fine-grained emotional variation in speech through controlled perturbation experiments, fine-grained emotion elicitation tasks, and representation geometry analysis. Our results show that vocal delivery variations systematically influence both model behavior and hidden representations, indicating that the evaluated LALMs already encode non-trivial speech-conditioned affective information. However, models struggle under emotionally ambiguous settings, particularly for semantically related fine-grained emotions. Representation analysis further suggests that subtle emotional structure is partially preserved despite weak downstream utilization. Overall, our findings suggest that the evaluated LALMs do not completely ignore subtle speech emotion cues, but instead utilize them inconsistently under ambiguous affective conditions

Relation prediction on knowledge graphs (KGs) attempts to infer the missing links between entities. Most previous studies are limited to the transductive setting where all entities must be seen during the training, making them unable to perform reasoning on emerging entities. Recently, the inductive setting is proposed to handle the entities in the test phase to be unseen during training, However, it suffers from the inefficient reasoning under the enclosing subgraph extraction issue and the lack of effective entity-independent feature modeling. To this end, we propose a novel Query Adaptive Anchor Representation (QAAR) model for inductive relation prediction. First, we extract one opening subgraph and perform reasoning by one time for all candidate triples, which is more efficient when the number of candidate triples is large. Second, we define some query adaptive anchors which are independent on any specific entity. Based on these anchors, we take advantage of the transferable entity-independent features (relation-aware, structure-aware and distance features) that can be used to produce entity embeddings for emerging unseen entities. Such entity-independent features is modeled by a query-aware graph attention network on the opening subgraph. Experimental results demonstrate that our proposed QAAR outperforms state-of-the-art baselines in inductive relation prediction task.

Cross-lingual entity alignment, which aims to match equivalent entities in KGs with different languages, has attracted considerable focus in recent years. Recently, many graph neural network (GNN) based methods are proposed for entity alignment and obtain promising results. However, existing GNN-based methods consider the two KGs independently and learn embeddings for different KGs separately, which ignore the useful pre-aligned links between two KGs. In this paper, we propose a novel Contextual Alignment Enhanced Cross Graph Attention Network (CAECGAT) for the task of cross-lingual entity alignment, which is able to jointly learn the embeddings in different KGs by propagating cross-KG information through pre-aligned seed alignments. We conduct extensive experiments on three benchmark cross-lingual entity alignment datasets. The experimental results demonstrate that our proposed method obtains remarkable performance gains compared to state-of-the-art methods.

The goal of Knowledge graph embedding (KGE) is to learn how to represent the low dimensional vectors for entities and relations based on the observed triples. The conventional shallow models are limited to their expressiveness. ConvE (Dettmers et al., 2018) takes advantage of CNN and improves the expressive power with parameter efficient operators by increasing the interactions between head and relation embeddings. However, there is no structural information in the embedding space of ConvE, and the performance is still limited by the number of interactions. The recent KBGAT (Nathani et al., 2019) provides another way to learn embeddings by adaptively utilizing structural information. In this paper, we take the benefits of ConvE and KBGAT together and propose a Relation-aware Inception network with joint local-global structural information for knowledge graph Embedding (ReInceptionE). Specifically, we first explore the Inception network to learn query embedding, which aims to further increase the interactions between head and relation embeddings. Then, we propose to use a relation-aware attention mechanism to enrich the query embedding with the local neighborhood and global entity information. Experimental results on both WN18RR and FB15k-237 datasets demonstrate that ReInceptionE achieves competitive performance compared with state-of-the-art methods.