Papers by Zhiheng Zheng

20 papers
LoRAMoE: Alleviating World Knowledge Forgetting in Large Language Models via MoE-Style Plugin (2024.acl-long)

Copied to clipboard

Challenge: Experimental results show that, as the instruction data increases, LoRAMoE can significantly improve the ability to process downstream tasks, while maintaining the world knowledge stored in the LLM.
Approach: They propose a framework that introduces several low-rank adapters and integrates them by using a router network to freeze the backbone model and force a portion of LoRAs to focus on leveraging world knowledge to solve downstream tasks.
Outcome: The proposed framework freezes the backbone model and forces a portion of LoRAs to focus on leveraging world knowledge to solve downstream tasks, to alleviate world knowledge forgetting.
Improving Discriminative Capability of Reward Models in RLHF Using Contrastive Learning (2024.emnlp-main)

Copied to clipboard

Challenge: Current methods rely on ranking losses to teach reward model to assess preferences, but they are susceptible to noise and ambiguous data, often failing to deeply understand human intentions.
Approach: They propose a method that incorporates contrastive learning into the reward modeling process to enhance generalization and stabilize the reinforcement learning training process.
Outcome: The proposed method enhances generalization of the reward model, stabilizes the reinforcement learning training process, and improves the final alignment with human preferences.
CityEQA: A Hierarchical LLM Agent on Embodied Question Answering Benchmark in City Space (2025.emnlp-main)

Copied to clipboard

Challenge: Embodied Question Answering (EQA) tasks are primarily focused on indoor environments, leaving the complexities of urban settings unexplored.
Approach: They propose a task where an embodied agent answers open-vocabulary questions in dynamic city spaces.
Outcome: The proposed agent achieves 60.7% of human-level answering accuracy compared to baselines . the proposed agent outperforms existing agents in open-ended city spaces .
Self-Polish: Enhance Reasoning in Large Language Models via Problem Refinement (2023.findings-emnlp)

Copied to clipboard

Challenge: Existing prompting methods have been used to enhance multistep reasoning capabilities of large language models, but they have overlooked the potential of formulating higher-quality problems.
Approach: They propose a method that starts from the problem side and refines problems to be more comprehensible and solvable for models.
Outcome: The proposed method achieves notable and consistent effectiveness on five reasoning benchmarks across different models.
AgentGym: Evaluating and Training Large Language Model-based Agents across Diverse Environments (2025.acl-long)

Copied to clipboard

Challenge: Large language models (LLMs) are promising foundations to build generally-capable agents . however, the community lacks a unified interactive framework that covers diverse environments for comprehensive evaluation of agents.
Approach: They propose a framework that features 7 real-world scenarios, 14 environments, and 89 tasks for unified, real-time, and concurrent agent interaction.
Outcome: The proposed framework features 7 real-world scenarios, 14 environments, and 89 tasks for unified, real-time, and concurrent agent interaction.
Multi-Programming Language Sandbox for LLMs (2025.acl-demo)

Copied to clipboard

Challenge: MPLSandbox is an out-of-the-box multi-programming language sandbox designed to provide unified and comprehensive feedback from compiler and analysis tools for Large Language Models (LLMs).
Approach: They propose a multi-programming language sandbox that provides unified feedback from compilers and analysis tools for Large Language Models.
Outcome: The proposed multi-language sandbox can provide comprehensive feedback from compilers and analysis tools for large language models (LLMs).
A Parallelized Framework for Simulating Large-Scale LLM Agents with Realistic Environments and Interactions (2025.acl-industry)

Copied to clipboard

Challenge: Existing work on large language models lacks a realistic environment and parallelized framework to support complex interactions between agents and environments.
Approach: They propose a framework that integrates realistic societal environments and parallelized interactions to support simulations of large-scale agents.
Outcome: The proposed framework can support simulations of 30,000 agents faster than the wall-clock time with 24 NVIDIA A800 GPUs and the performance increases linearly with the increase of LLM computational resources.
Subspace Defense: Discarding Adversarial Perturbations by Learning a Subspace for Clean Signals (2024.lrec-main)

Copied to clipboard

Challenge: Existing models that extract discrete inputs into fixed-length representations are vulnerable to adversarial attacks that place perturbations on clean inputs to fool DNNs.
Approach: They propose to inspect the subspaces of sample features through spectral analysis to better understand adversarial attacks.
Outcome: The proposed strategy enables the model to inherently suppress adversaries, which boosts model robustness and motivates new directions of effective adversarial defense.
SWE-QA-Pro: A Representative Benchmark and Scalable Training Recipe for Repository-Level Code Understanding (2026.findings-acl)

Copied to clipboard

Challenge: Existing benchmarks for agentic repository-level code understanding overlook long tail topics and rely on memorized knowledge.
Approach: They propose a repository-level agentic code understanding benchmark that uses long-tail repositories with executable environments to enforce topical balance.
Outcome: Empirically, a Qwen3-8B model trained with the proposed benchmark outperforms GPT-4o by 2.3 points.
Connectivity Patterns are Task Embeddings (2023.findings-acl)

Copied to clipboard

Challenge: Existing methods for predicting inter-task transferability are sparse and task-specific.
Approach: They propose a method that uses connectivity patterns of neurons as a unique identifier associated with a task.
Outcome: The proposed method outperforms baselines in predicting inter-task transferability across data regimes and transfer settings while keeping high efficiency in computation and storage.
Reward Modeling Requires Automatic Adjustment Based on Data Quality (2024.findings-emnlp)

Copied to clipboard

Challenge: Reinforcement Learning from Human Feedback (RLHF) is a method for aligning language models with human values.
Approach: They propose a method that automatically adjusts reward modeling based on data quality . they use preference data to train a reward model that is more aligned with human values .
Outcome: The proposed method stabilizes reward model training and significantly improves alignment performance on human preference datasets.
Characterizing the Impacts of Instances on Robustness (2023.findings-acl)

Copied to clipboard

Challenge: Existing defense approaches focus on developing new model structures or training algorithms, but they do little to tap the potential of training instances.
Approach: They propose a method that can distinguish between robust and non-robust instances according to the model’s sensitivity to perturbations on individual instances during training.
Outcome: The proposed method can distinguish between robust and non-robust instances according to the model’s sensitivity to perturbations on individual instances during training.
ORTicket: Let One Robust BERT Ticket Transfer across Different Tasks (2024.lrec-main)

Copied to clipboard

Challenge: Pretrained language models are susceptible to subtle perturbations and require multiple adversarial training during fine-tuning to improve their robustness.
Approach: They propose a novel adversarial defense method ORTicket that fine-tunes a model for downstream tasks.
Outcome: The proposed method achieves comparable robustness to other defense methods while maintaining the efficiency of fine-tuning.
Toward Optimal LLM Alignments Using Two-Player Games (2025.findings-emnlp)

Copied to clipboard

Challenge: Alignment of large language models (LLM) is a process that ensures the model’s responses to user prompts align with human intentions and social values.
Approach: They propose an alignment method based on a two-agent game consisting of an adversarial agent and a defensive agent.
Outcome: The proposed method improves on a two-agent game with an adversarial agent and a defensive agent.
Efficient Adversarial Training with Robust Early-Bird Tickets (2022.emnlp-main)

Copied to clipboard

Challenge: Existing methods to improve the robustness of pre-trained language models are expensive because of the need to generate adversarial examples via gradient descent.
Approach: They propose an adversarial optimization method that searches for robust tickets with structured sparsity in the early stage and fine-tunes tickets in the remaining time.
Outcome: The proposed method achieves up to 7 13 training speedups while maintaining comparable or even better robustness compared to the most competitive state-of-the-art methods.
ABC-Bench: Benchmarking Agentic Backend Coding in Real-World Development (2026.findings-acl)

Copied to clipboard

Challenge: Large Language Models (LLMs) have redefined the role of AI in software engineering . current benchmarks focus on localized code generation, but neglect dynamic, full-process requirements of real-world engineering.
Approach: They propose a benchmark to evaluate agentic backend coding within a realistic, executable workflow.
Outcome: The ABC-Bench benchmark evaluates agentic backend coding within a realistic, executable workflow.
StepCoder: Improving Code Generation with Reinforcement Learning from Compiler Feedback (2024.acl-long)

Copied to clipboard

Challenge: Existing work integrates reinforcement learning with compiler feedback to enhance code generation quality but the long code generated by LLMs makes RL exploration ineffective.
Approach: They propose a framework that integrates reinforcement learning and compiler feedback to enhance code generation quality.
Outcome: The proposed framework outperforms state-of-the-art approaches in corresponding benchmarks and integrates reinforcement learning with compiler feedback to improve code generation quality.
RealBehavior: A Framework for Faithfully Characterizing Foundation Models’ Human-like Behavior Mechanisms (2023.findings-emnlp)

Copied to clipboard

Challenge: Existing studies on human-like behaviors in foundation models do not verify their faithfulness . a simple application of psychological tools cannot faithfully characterize all human-type behaviors .
Approach: They propose a framework to characterize humanoid behaviors in foundation models . they argue that a simple application of psychological tools cannot faithfully characterize all human-like behaviors .
Outcome: The proposed framework assesses the faithfulness of results based on reproducibility, internal consistency, and generalizability.
Mitigating Object Hallucinations in MLLMs via Multi-Frequency Perturbations (2025.findings-emnlp)

Copied to clipboard

Challenge: Multimodal large language models have demonstrated remarkable performance in visual-language tasks, but their authenticity is often compromised by object hallucinations.
Approach: They propose a multi-frequency perturbation method that leverages both low-frequency and high-frequency features of images to perturb visual feature representations and explicitly suppress redundant frequency-domain features during inference.
Outcome: The proposed method significantly mitigates object hallucinations across various model architectures.
TextFlint: Unified Multilingual Robustness Evaluation Toolkit for Natural Language Processing (2021.acl-demo)

Copied to clipboard

Challenge: Existing approaches to textual robustness evaluation focus on slightly modifying the input data, which maintains the original meaning and results in a different prediction.
Approach: They propose a multilingual robustness evaluation toolkit for NLP that integrates universal text transformations, task-specific transformations and adversarial attack.
Outcome: The toolkit includes universal text transformation, task-specific transformation, adversarial attack, subpopulation, and their combinations to provide comprehensive robustness analyses.

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