Papers by Jiahui Zhang

21 papers
Meta-rater: A Multi-dimensional Data Selection Method for Pre-training Language Models (2025.acl-long)

Copied to clipboard

Challenge: composition of pre-training datasets for large language models remains undisclosed . current methods for evaluating data quality are limited by single-dimensional evaluation or redundancy-focused strategies.
Approach: They propose a multi-dimensional data selection method that integrates dimensions with existing quality metrics through learned optimal weightings.
Outcome: The proposed method doubles convergence speed for 1.3B model models and improves downstream task performance by 3.23%.
DetGPT: Detect What You Need via Reasoning (2023.emnlp-main)

Copied to clipboard

Challenge: Recent advances in the field of computer vision have enabled more effective and sophisticated interactions between humans and machines.
Approach: They propose a reasoning-based object detection paradigm that leverages state-of-the-art multi-modal models and open-vocabulary object detectors to perform reasoning within the context of the user’s instructions and the visual scene.
Outcome: The proposed method enables users to interact with the system using natural language instructions, allowing for a higher level of interactivity.
Exploring Multimodal Relation Extraction of Hierarchical Tabular Data with Multi-task Learning (2025.acl-long)

Copied to clipboard

Challenge: Existing studies overlook the need of mining relations among multiple columns rather than just the semantic relation between two specific columns in real-world practice.
Approach: They propose a Chain-of-Thought distillation framework with self-correction mechanism to enhance MLLMs’ reasoning capabilities without increasing parameter scale.
Outcome: The proposed method significantly outperforms baselines on wide datasets.
SGPVT: Self-Generated Proximal Visual Tokens for Mitigating Proximal Collateral Damage in MLLM Unlearning (2026.acl-long)

Copied to clipboard

Challenge: Existing approaches focus on general utility metrics, overlooking the preservation of semantically related concepts.
Approach: They propose a method that introduces self-generated proximal visual tokens to prevent forgetting vulnerability.
Outcome: The proposed framework outperforms existing methods in preserving semantically related concepts while achieving effective target unlearning.
Look Less, Reason More: Rollout-Guided Adaptive Pixel-Space Reasoning (2026.acl-long)

Copied to clipboard

Challenge: Recent work has shown promise by incorporating pixel-level visual information into the reasoning process, enabling VLMs to access high-resolution visual details during their thought process.
Approach: They propose a framework that dynamically determines necessary pixel-level operations based on the input query.
Outcome: The proposed model achieves 73.4% accuracy on HR-Bench 4K while maintaining a tool usage ratio of only 20.1%, improving accuracy and reducing tool usage by 66.5% compared to the previous methods.
VocalRep: Structure-Aware Vocal Representations for Multimodal Generation (2026.findings-acl)

Copied to clipboard

Challenge: Existing approaches to vocal separation are optimized for signal-level reconstruction, but they overlook structural disentanglement required for downstream generation tasks.
Approach: They propose a structure-aware learning framework to disentangle vocals, harmonies, and accompaniment . they combine global vocal identity conditioning with ranking-based objectives .
Outcome: The proposed framework disentangles lead vocals, harmonies, and accompaniment while enforcing role consistency across long-form audio.
Efficient Pretraining Data Selection for Language Models via Multi-Actor Collaboration (2025.acl-long)

Copied to clipboard

Challenge: Efficient data selection is crucial to accelerate the pretraining of language models . limited research has addressed the inherent conflicts between data selection methods .
Approach: They propose a multi-actor collaborative data selection mechanism that prioritizes data based on its specific criterion and updates prioritization rules using the current state of the model.
Outcome: The proposed model accelerates convergence in LM pretraining and achieves an average relative performance gain of 10.5% across multiple language model benchmarks.
Reasoning over Precedents Alongside Statutes: Case-Augmented Deliberative Alignment for LLM Safety (2026.acl-long)

Copied to clipboard

Challenge: OpenAI introduces deliberative alignment (DA) to enhance safety of its o-series models, but effectiveness of this approach in open-source LLMs is understudied.
Approach: They propose a case-augmented deliberative alignment method for large language models . they propose to use reinforcement learning on self-generated safety reasoning chains .
Outcome: The proposed method avoids narrowly enumerated rules and allows broader adaptability.
Marco-Bench-MIF: On Multilingual Instruction-Following Capability of Large Language (2025.acl-long)

Copied to clipboard

Challenge: Existing datasets for instruction-following are monolingual and centered on English . existing data are unable to capture linguistic and cultural subtle differences .
Approach: They propose an extension of IFEval to a localized multilingual version called Marco-Bench-MIF . their benchmark addresses linguistic constraints and cultural references via translation and verification .
Outcome: The proposed extension of IFEval to a localized multilingual version covers 30 languages with varying levels of localization.
MultiPL-MoE: Multi-Programming-Lingual Extension of Large Language Models through Hybrid Mixture-of-Experts (2025.findings-emnlp)

Copied to clipboard

Challenge: MultiPL is a special case of multiple natural languages and requires limited computational resources to generate multilingual code.
Approach: They propose to extend LLMs by combining two paired experts to optimize expert selection at token and segment levels.
Outcome: The proposed extension improves the performance of the base LLMs while retaining the most popular ones using limited computational resources.
Think Faster Than Words: Efficient LLM Chain-of-Thought Reasoning via Dynamic Shortcut Decoding (2026.acl-long)

Copied to clipboard

Challenge: Existing methods that prune or employ early stopping to reduce latency often compromise reasoning reliability.
Approach: They propose a shortcut decoding framework that integrates probes over internal hidden states with step-level entropy to detect convergence of reasoning during generation and adaptively selects between a fast-exit path and a stability-verified path to remove redundant steps while preserving answer correctness.
Outcome: The proposed framework reduces token usage by approximately 35% and maintains accuracy comparable to full CoT decoding.
Pointing to a Llama and Call it a Camel: On the Sycophancy of Multimodal Large Language Models (2025.emnlp-main)

Copied to clipboard

Challenge: Multimodal large language models exhibit a pronounced form of visual sycophantic behavior when they process image inputs.
Approach: They propose a technique that allows multimodal large language models to engage in reflective reasoning and determine whether a user’s instruction is misleading or corrective.
Outcome: The proposed model resists misleading instructions but is stubborn even if it is wrong.
Optimizing Language Models with Fair and Stable Reward Composition in Reinforcement Learning (2024.emnlp-main)

Copied to clipboard

Challenge: Recent research has developed algorithms for reinforcement learning from human feedback and AI-generated feedback.
Approach: They propose a method for reinforcement learning from human feedback and AI-generated feedback that incorporates weighting, ranking, and constraining to handle disparate rewards.
Outcome: The proposed method reduces disparity and enhances stability among rewards . empirical results show that the proposed method is efficient and straightforward .
Don’t Be Misled by Style: A Style-Adaptive Reranker for Capturing Effective Knowledge in Retrieval-Augmented Generation (2026.acl-long)

Copied to clipboard

Challenge: Existing rerankers are mainly trained on well-edited texts, but stylistic features can be misled by reranked models.
Approach: They propose a style-augmented multi-task framework that prioritizes effective knowledge over stylistic perturbations by using an LLM to derive passage-level supervision on whether a passage helps or harms answer correctness.
Outcome: Extensive experiments show that SARK improves generation performance across multiple LLMs under mixed-style conditions.
Unveiling Opinion Evolution via Prompting and Diffusion for Short Video Fake News Detection (2024.findings-acl)

Copied to clipboard

Challenge: Existing methods for short video fake news detection ignore the implicit opinions and evolving nature of opinions across modalities.
Approach: They propose a short video fake news model that mines implicit opinions within short videos and promotes the evolution of both explicit and implicit opinions across all modalities.
Outcome: The proposed model outperforms existing methods on a publicly available dataset for short video fake news detection.
ExeSQL: Self-Taught Text-to-SQL Models with Execution-Driven Bootstrapping for SQL Dialects (2025.findings-emnlp)

Copied to clipboard

Challenge: Existing text-to-SQL models are limited to SQLite due to dataset limitations . data generated through static prompting is noisy and unreliable, authors say .
Approach: They propose a text-to-SQL framework with execution-driven, agentic bootstrapping . ExeSQl bridges the dialect gap in text- to-Sql, achieving average improvements .
Outcome: ExeSQL bridges the dialect gap in text-to-SQl, with average improvements of 15.2%, 10.38%, and 4.49% over GPT-4o on PostgreSQLE, MySQL, and Oracle.
Shaping the Safety Boundaries: Understanding and Defending Against Jailbreaks in Large Language Models (2025.acl-long)

Copied to clipboard

Challenge: Understanding how jailbreaking works remains limited, hindering the development of effective defense strategies.
Approach: They propose a new mechanism that adaptively constrains activations within the safety boundary and propose 'Activation Boundary Defense' to enhance its effectiveness.
Outcome: The proposed defense achieves an average Defense Success Rate (DSR) of over 98% against various jailbreak attacks, with less than 2% impact on the model’s general capabilities.
From Discrimination to Generation: Low-Resource Intent Detection with Language Model Instruction Tuning (2024.findings-acl)

Copied to clipboard

Challenge: Existing studies fine-tune discriminative models on specific defined intent classes, preventing them from being directly adopted to new intent domains.
Approach: They propose to use a pre-trained generative intent model to detect new intents from different domains with no parameter updates.
Outcome: The proposed model outperforms baselines that need further fine-tuning or domain-specific samples.
Clear Up Confusion: Iterative Differential Generation for Fine-grained Intent Detection with Contrastive Feedback (2025.coling-main)

Copied to clipboard

Challenge: Recent studies on fine-grained intent detection have focused on collecting large-scale and high-quality samples via crowdsourcing resulting in data scarcity.
Approach: They propose an iterative differential generation framework with contrastive feedback to generate high-quality pseudo samples and accurately capture the crucial nuances in target class distribution.
Outcome: The proposed framework generates high-quality pseudo samples and captures crucial nuances in target class distribution.
Mixture of insighTful Experts (MoTE): The Synergy of Reasoning Chains and Expert Mixtures in Self-Alignment (2025.acl-long)

Copied to clipboard

Challenge: Recent studies show that reasoning abilities contribute significantly to model safety, while integrating Mixture-of-Experts (MoE) architectures can further enhance alignment.
Approach: They propose a framework that synergistically combines reasoning chains and expert mixtures to improve self-alignment.
Outcome: The proposed framework improves model safety, jailbreak resistance, and over-refusal capabilities, achieving performance comparable to OpenAI’s state-of-the-art o1 model.
Awakening Latent Grounding from Pretrained Language Models for Semantic Parsing (2021.findings-acl)

Copied to clipboard

Challenge: Recent years pretrained language models (PLMs) have shown their power on modeling language . however, few efforts have been made to explore grounding capabilities of PLMs .
Approach: They propose to use pretrained language models to explore syntactic structures . they propose to combine their approach with an erasingthen-awakening approach . their results show that the approach can awaken latent grounding, which is understandable to humans .
Outcome: Empirical studies show that the proposed approach can awaken latent grounding . it shows great potential to benefit downstream semantic parsing models, it says .

What is GenGO?

GenGO is an NLP powered publication search system. It currenctly indexes 30k+ papers from ACL Anthology, and implements multi-aspect summarization, semantic search, and more!

Information

About
Limitations