LLMs have immense potential for generating plans, transforming an initial world state into a desired goal state. A large body of research has explored the use of LLMs for various planning tasks, from web navigation to travel planning and database querying. However, many of these systems are tailored to specific problems, making it challenging to compare them or determine the best approach for new tasks. There is also a lack of clear and consistent evaluation criteria. Our survey aims to offer a comprehensive overview of current LLM planners to fill this gap. It builds on foundational work by Kartam and Wilkins (1990) and examines six key performance criteria: completeness, executability, optimality, representation, generalization, and efficiency. For each, we provide a thorough analysis of representative works and highlight their strengths and weaknesses. Our paper also identifies crucial future directions, making it a valuable resource for both practitioners and newcomers interested in leveraging LLM planning to support agentic workflows.

Named entity recognition is a fundamental task in ancient Chinese text analysis.Based on the pre-trained language model of ancient Chinese texts, this paper proposes a new named entity recognition method GRoWE. It uses the ancient Chinese texts pre-trained language model GujiRoBERTa as the base model, and the wordword relation prediction model is superposed upon the base model to construct a superposition model. Then ensemble strategies are used to multiple superposition models. On the EvaHan 2025 public test set, the F1 value of the proposed method reaches 86.79%, which is 6.18% higher than that of the mainstream BERT_LSTM_CRF baseline model, indicating that the model architecture and ensemble strategy play an important role in improving the recognition effect of naming entities in ancient Chinese texts.

Recent advances in Large Language Models (LLMs) have enabled them to process increasingly longer sequences, ranging from 2K to 2M tokens and even beyond. However, simply extending the input sequence length does not necessarily lead to effective long-context understanding. In this study, we integrate Chain-of-Thought (CoT) reasoning into LLMs in a supervised manner to facilitate effective long-context understanding. To achieve this, we introduce LongFinanceQA, a synthetic dataset in the financial domain designed to improve long-context reasoning. Unlike existing long-context synthetic data, LongFinanceQA includes intermediate CoT reasoning before the final conclusion, which encourages LLMs to perform explicit reasoning, improving accuracy and interpretability in long-context understanding. To generate synthetic CoT reasoning, we propose Property-based Agentic Inference (PAI), an agentic framework that simulates human-like reasoning steps, including property extraction, retrieval, and summarization. We evaluate PAI’s reasoning capabilities by assessing GPT-4o-mini w/ PAI on the Loong benchmark, outperforming standard GPT-4o-mini by 20.0%. Furthermore, we fine-tune LLaMA-3.1-8B-Instruct on LongFinanceQA, achieving a 28.0% gain on Loong’s financial subset.

Supervised fine-tuning (SFT) on benign data can paradoxically erode a language model’s safety alignment, a phenomenon known as catastrophic forgetting of safety behaviors. Although prior work shows that randomly adding safety examples can reduce harmful output, the principles that make certain examples more effective than others remain poorly understood. This paper investigates the hypothesis that the effectiveness of a safety example is governed by two key factors: its instruction-response behavior (e.g., refusal vs. explanation) and its semantic diversity across harm categories. We systematically evaluate sampling strategies based on these axes and find that structured, diversity-aware sampling significantly improves model safety. Our method reduces harmfulness by up to 41% while adding only 0.05% more data to the fine-tuning set.

Bargaining is an important and unique part of negotiation between humans. As LLM-driven agents learn to negotiate and act like real humans, how to evaluate agents’ bargaining abilities remains an open problem.For the first time, we formally described the Bargaining task as an asymmetric incomplete information game, defining the gains of the Buyer and Seller in multiple bargaining processes. It allows us to quantitatively assess an agent’s performance in the Bargain task.We collected a real product price dataset, AmazonHistoryPrice, and conducted evaluations of various LLM agents’ bargaining abilities. We find that playing a Buyer is much harder than a Seller, and increasing model size can not effectively improve the Buyer’s performance.To address the challenge, we propose a novel approach called OG-Narrator that integrates a deterministic Offer Generator to control the price range of Buyer’s offers, and an LLM Narrator to create natural language sentences for generated offers.Experimental results show that OG-Narrator improves the buyer’s deal rates from 26.67% to 88.88% and brings a ten times multiplication of profits on all baselines, even a model that has not been aligned.

Large language models (LLMs) have exhibited great potential in autonomously completing tasks across real-world applications. Despite this, these LLM agents introduce unexpected safety risks when operating in interactive environments. Instead of centering on the harmlessness of LLM-generated content in most prior studies, this work addresses the imperative need for benchmarking the behavioral safety of LLM agents within diverse environments. We introduce R-Judge, a benchmark crafted to evaluate the proficiency of LLMs in judging and identifying safety risks given agent interaction records. R-Judge comprises 569 records of multi-turn agent interaction, encompassing 27 key risk scenarios among 5 application categories and 10 risk types. It is of high-quality curation with annotated safety labels and risk descriptions. Evaluation of 11 LLMs on R-Judge shows considerable room for enhancing the risk awareness of LLMs: The best-performing model, GPT-4o, achieves 74.42% while no other models significantly exceed the random. Moreover, we reveal that risk awareness in open agent scenarios is a multi-dimensional capability involving knowledge and reasoning, thus challenging for LLMs. With further experiments, we find that fine-tuning on safety judgment significantly improve model performance while straightforward prompting mechanisms fail. R-Judge is publicly available at Annoymous.

The SPEADO model for sentence segmentation and punctuation tasks in ancient Chinese texts is proposed, which incorporates text chunking and MinHash indexing techniques to realise example argumentation. Additionally, decoding optimization strategies are introduced to direct the attention of the LLM model towards punctuation errors and address the issue of uncontrollable output. Experimental results show that the F1 score of the proposed method exceeds the baseline model by 14.18%, indicating a significant improvement in performance.

Matching equivalent entities across Knowledge graphs is a pivotal step for knowledge fusion. Previous approaches usually study the problem in Euclidean space. However, recent works have shown that hyperbolic space has a higher capacity than Euclidean space and hyperbolic embedding can represent the hierarchical structure in a knowledge graph. In this paper, we propose a localized geometric method to find equivalent entities in hyperbolic space. Specifically, we use a hyperbolic neural network to encode the lingual information of entities and the structure of both knowledge graphs into a low-dimensional hyperbolic space. To address the asymmetry of structure on different KGs and the localized nature of relations, we learn an instance-specific geometric mapping function based on rotation to match entity pairs. A contrastive loss function is used to train the model. The experiment verifies the power of low-dimensional hyperbolic space for entity matching and shows that our method outperforms the state of the art by a large margin.

Disease is one of the fundamental entities in biomedical research. Recognizing such entities from biomedical text and then normalizing them to a standardized disease vocabulary offer a tremendous opportunity for many downstream applications. Previous studies have demonstrated that joint modeling of the two sub-tasks has superior performance than the pipelined counterpart. Although the neural joint model based on multi-task learning framework has achieved state-of-the-art performance, it suffers from the boundary inconsistency problem due to the separate decoding procedures. Moreover, it ignores the rich information (e.g., the text surface form) of each candidate concept in the vocabulary, which is quite essential for entity normalization. In this work, we propose a neural transition-based joint model to alleviate these two issues. We transform the end-to-end disease recognition and normalization task as an action sequence prediction task, which not only jointly learns the model with shared representations of the input, but also jointly searches the output by state transitions in one search space. Moreover, we introduce attention mechanisms to take advantage of the text surface form of each candidate concept for better normalization performance. Experimental results conducted on two publicly available datasets show the effectiveness of the proposed method.

This paper describes our system developed for the subtask 1c of the sixth Social Media Mining for Health Applications (SMM4H) shared task in 2021. The aim of the subtask is to recognize the adverse drug effect (ADE) mentions from tweets and normalize the identified mentions to their mapping MedDRA preferred term IDs. Our system is based on a neural transition-based joint model, which is to perform recognition and normalization simultaneously. Our final two submissions outperform the average F1 score by 1-2%.

Marge infusé algorithmes détendus (MIRAS) dominent modèle de tuning dans la traduction automatique statistique dans le cas des grandes caractéristiques de l’échelle, mais ils sont également célèbres pour la complexité de mise en œuvre. Nous introduisons une nouvelle méthode, qui concerne une liste des N meilleures comme une permutation et minimise la perte Plackett-Luce de permutations rez-de-vérité. Des expériences avec des caractéristiques à grande échelle démontrent que, la nouvelle méthode est plus robuste que MERT ; si ce est seulement à rattacher avec Miras, il a un avantage comparativement, plus facile à mettre en œuvre.

This paper describes the ICT Statistical Machine Translation systems that used in the evaluation campaign of the International Workshop on Spoken Language Translation (IWSLT) 2009. For this year’s evaluation, we participated in the Challenge Task (Chinese-English and English-Chinese) and BTEC Task (Chinese-English). And we mainly focus on one new method to improve single system’s translation quality. Specifically, we developed a sentence-similarity based development set selection technique. For each task, we finally submitted the single system who got the maximum BLEU scores on the selected development set. The four single translation systems are based on different techniques: a linguistically syntax-based system, two formally syntax-based systems and a phrase-based system. Typically, we didn’t use any rescoring or system combination techniques in this year’s evaluation.