Exploring Model Editing for LLM-based Aspect-Based Sentiment Classification

Model editing aims at selectively updating a small subset of a neural model's parameters with an interpretable strategy to achieve desired modifications. It can significantly reduce computational costs to adapt to large language models (LLMs). Given its ability to precisely target critical components within LLMs, model editing shows great potential for efficient fine-tuning applications. In this work, we investigate model editing to serve an efficient method for adapting LLMs to solve aspect-based sentiment classification. Through causal interventions, we trace and determine which neuron hidden states are essential for the prediction of the model. By performing interventions and restorations on each component of an LLM, we identify the importance of these components for aspect-based sentiment classification. Our findings reveal that a distinct set of mid-layer representations is essential for detecting the sentiment polarity of given aspect words. Leveraging these insights, we develop a model editing approach that focuses exclusively on these critical parts of the LLM, leading to a more efficient method for adapting LLMs. Our in-domain and out-of-domain experiments demonstrate that this approach achieves competitive results compared to the currently strongest methods with significantly fewer trainable parameters, highlighting a more efficient and interpretable fine-tuning strategy.

DynaCode: A Dynamic Complexity-Aware Code Benchmark for Evaluating Large Language Models in Code Generation

The rapid advancement of large language models (LLMs) has significantly improved their performance in code generation tasks. However, existing code benchmarks remain static, consisting of fixed datasets with predefined problems. This makes them vulnerable to memorization during training, where LLMs recall specific test cases instead of generalizing to new problems, leading to data contamination and unreliable evaluation results. To address these issues, we introduce DynaCode, a dynamic, complexity-aware benchmark that overcomes the limitations of static datasets. DynaCode evaluates LLMs systematically using a complexity-aware metric, incorporating both code complexity and call-graph structures. DynaCode achieves large-scale diversity, generating up to 189 million unique nested code problems across four distinct levels of code complexity, referred to as units, and 16 types of call graphs. Results on 12 latest LLMs show an average performance drop of 16.8% to 45.7% compared to MBPP+, a static code generation benchmark, with performance progressively decreasing as complexity increases. This demonstrates DynaCode's ability to effectively differentiate LLMs. Additionally, by leveraging call graphs, we gain insights into LLM behavior, particularly their preference for handling subfunction interactions within nested code.

DeepReview: Improving LLM-based Paper Review with Human-like Deep Thinking Process

Large Language Models (LLMs) are increasingly utilized in scientific research assessment, particularly in automated paper review. However, existing LLM-based review systems face significant challenges, including limited domain expertise, hallucinated reasoning, and a lack of structured evaluation. To address these limitations, we introduce DeepReview, a multi-stage framework designed to emulate expert reviewers by incorporating structured analysis, literature retrieval, and evidence-based argumentation. Using DeepReview-13K, a curated dataset with structured annotations, we train DeepReviewer-14B, which outperforms CycleReviewer-70B with fewer tokens. In its best mode, DeepReviewer-14B achieves win rates of 88.21\% and 80.20\% against GPT-o1 and DeepSeek-R1 in evaluations. Our work sets a new benchmark for LLM-based paper review, with all resources publicly available. The code, model, dataset and demo have be released in http://ai-researcher.net.

Cell-Free MIMO-ISAC: Joint Location and Velocity Estimation and Fundamental CRLB Analysis

This paper investigates joint location and velocity estimation, along with their fundamental performance bounds analysis, in a cell-free multi-input multi-output (MIMO) integrated sensing and communication (ISAC) system. First, unlike existing studies that derive likelihood functions for target parameter estimation using continuous received signals, we formulate the maximum likelihood estimation (MLE) for radar sensing based on discrete received signals at a given sampling rate. Second, leveraging the proposed MLEs, we derive closed-form Cramer-Rao lower bounds (CRLBs) for joint location and velocity estimation in both single-target and multiple-target scenarios. Third, to enhance computational efficiency, we propose approximate CRLBs and conduct an in-depth accuracy analysis. Additionally, we thoroughly examine the impact of sampling rate, squared effective bandwidth, and time width on CRLB performance. For multiple-target scenarios, the concepts of safety distance and safety velocity are introduced to characterize conditions under which the CRLBs for multiple targets converge to their single target counterparts. Finally, extensive simulations are conducted to verify the accuracy of the proposed CRLBs and the theoretical results using state-of-the-art waveforms, namely orthogonal frequency division multiplexing (OFDM) and orthogonal chirp division multiplexing (OCDM).

Joint Beamforming and Compressed Sensing for Uplink Grant-Free Access

Compressed sensing (CS)-based techniques have been widely applied in the grant-free non-orthogonal multiple access (NOMA) to a single-antenna base station (BS). In this paper, we consider the multi-antenna reception at the BS for uplink grant-free access for the massive machine type communication (mMTC) with limited channel resources. To enhance the overloading performance of the BS, we develop a general framework for the synergistic amalgamation of the spatial division multiple access (SDMA) technique with the CS-based grant-free NOMA. We derive a closed-form statistical beamforming and a dynamic beamforming scheme for the inter-cluster interference suppression when applying SDMA. Based on this, we further develop a joint adaptive beamforming and subspace pursuit (JABF-SP) algorithm for the multiuser detection and data recovery, with a novel sparsity level decision method without the accurate knowledge of the noise level. To further improve the data recovery performance, we propose an interference cancellation based J-ABF-SP scheme (J-ABF-SP-IC) by using the initial signal estimates generated from the J-ABF-SP algorithm. Illustrative simulations verify the superior user detection and signal recovery performance of our proposed algorithms in comparison with existing CS-based grant-free NOMA techniques.

Lost in Literalism: How Supervised Training Shapes Translationese in LLMs

Large language models (LLMs) have achieved remarkable success in machine translation, demonstrating impressive performance across diverse languages. However, translationese, characterized by overly literal and unnatural translations, remains a persistent challenge in LLM-based translation systems. Despite their pre-training on vast corpora of natural utterances, LLMs exhibit translationese errors and generate unexpected unnatural translations, stemming from biases introduced during supervised fine-tuning (SFT). In this work, we systematically evaluate the prevalence of translationese in LLM-generated translations and investigate its roots during supervised training. We introduce methods to mitigate these biases, including polishing golden references and filtering unnatural training instances. Empirical evaluations demonstrate that these approaches significantly reduce translationese while improving translation naturalness, validated by human evaluations and automatic metrics. Our findings highlight the need for training-aware adjustments to optimize LLM translation outputs, paving the way for more fluent and target-language-consistent translations. We release the data and code at https://github.com/yafuly/LLM_Translationese.

LAG: LLM agents for Leaderboard Auto Generation on Demanding

This paper introduces Leaderboard Auto Generation (LAG), a novel and well-organized framework for automatic generation of leaderboards on a given research topic in rapidly evolving fields like Artificial Intelligence (AI). Faced with a large number of AI papers updated daily, it becomes difficult for researchers to track every paper's proposed methods, experimental results, and settings, prompting the need for efficient automatic leaderboard construction. While large language models (LLMs) offer promise in automating this process, challenges such as multi-document summarization, leaderboard generation, and experiment fair comparison still remain under exploration. LAG solves these challenges through a systematic approach that involves the paper collection, experiment results extraction and integration, leaderboard generation, and quality evaluation. Our contributions include a comprehensive solution to the leaderboard construction problem, a reliable evaluation method, and experimental results showing the high quality of leaderboards.

RefuteBench 2.0 -- Agentic Benchmark for Dynamic Evaluation of LLM Responses to Refutation Instruction

In the multi-turn interaction schema, large language models (LLMs) can leverage user feedback to enhance the quality and relevance of their responses. However, evaluating an LLM's ability to incorporate user refutation feedback is crucial yet challenging. In this study, we introduce RefuteBench 2.0, which significantly extends the original RefuteBench by incorporating LLM agents as refuters and evaluators, which allows for flexible and comprehensive assessment. We design both transient and persistent refutation instructions with different validity periods. Meta-evaluation shows that the LLM-based refuter could generate more human-like refutations and the evaluators could assign scores with high correlation with humans. Experimental results of various LLMs show that current models could effectively satisfy the refutation but fail to memorize the refutation information. Interestingly, we also observe that the performance of the initial task decreases as the refutations increase. Analysis of the attention scores further shows a potential weakness of current LLMs: they struggle to retain and correctly use previous information during long context dialogues. https://github.com/ElliottYan/RefuteBench-2.0

Sketch-1-to-3: One Single Sketch to 3D Detailed Face Reconstruction

3D face reconstruction from a single sketch is a critical yet underexplored task with significant practical applications. The primary challenges stem from the substantial modality gap between 2D sketches and 3D facial structures, including: (1) accurately extracting facial keypoints from 2D sketches; (2) preserving diverse facial expressions and fine-grained texture details; and (3) training a high-performing model with limited data. In this paper, we propose Sketch-1-to-3, a novel framework for realistic 3D face reconstruction from a single sketch, to address these challenges. Specifically, we first introduce the Geometric Contour and Texture Detail (GCTD) module, which enhances the extraction of geometric contours and texture details from facial sketches. Additionally, we design a deep learning architecture with a domain adaptation module and a tailored loss function to align sketches with the 3D facial space, enabling high-fidelity expression and texture reconstruction. To facilitate evaluation and further research, we construct SketchFaces, a real hand-drawn facial sketch dataset, and Syn-SketchFaces, a synthetic facial sketch dataset. Extensive experiments demonstrate that Sketch-1-to-3 achieves state-of-the-art performance in sketch-based 3D face reconstruction.

ThinkBench: Dynamic Out-of-Distribution Evaluation for Robust LLM Reasoning

Evaluating large language models (LLMs) poses significant challenges, particularly due to issues of data contamination and the leakage of correct answers. To address these challenges, we introduce ThinkBench, a novel evaluation framework designed to evaluate LLMs' reasoning capability robustly. ThinkBench proposes a dynamic data generation method for constructing out-of-distribution (OOD) datasets and offers an OOD dataset that contains 2,912 samples drawn from reasoning tasks. ThinkBench unifies the evaluation of reasoning models and non-reasoning models. We evaluate 16 LLMs and 4 PRMs under identical experimental conditions and show that most of the LLMs' performance are far from robust and they face a certain level of data leakage. By dynamically generating OOD datasets, ThinkBench effectively provides a reliable evaluation of LLMs and reduces the impact of data contamination.