Papers by Jingfeng Zhang

9 papers
Hephaestus: Improving Fundamental Agent Capabilities of Large Language Models through Continual Pre-Training (2025.naacl-long)

Copied to clipboard

Challenge: Existing LLMs often rely on complex prompting or extensive fine-tuning to introduce new capabilities while preserving strong generalizability.
Approach: They propose a large-scale pre-training corpus to enhance LLM agents' capabilities . they use 103B agent-specific data encompassing 76,537 APIs .
Outcome: The proposed training corpus outperforms open-source LLMs and commercial LLM agents on three agent benchmarks.
Fix-Filter-Fix: Intuitively Connect Any Models for Effective Bug Fixing (2021.emnlp-main)

Copied to clipboard

Challenge: Existing approaches for bug fixing lack generality and use only textual or structured information.
Approach: They propose an intuitive yet effective general framework called Fix-Filter-Fix for bug fixing that connects models with their filter mechanism to filter out the last model’s unchanged fix to the next.
Outcome: The proposed framework can quantify and accurately calculate the lifting effect of the model.
LongLeader: A Comprehensive Leaderboard for Large Language Models in Long-context Scenarios (2025.naacl-long)

Copied to clipboard

Challenge: LongLeader aims to assess different LLMs' long-context comprehension abilities . long-constext comprehension is a key bottleneck for many use cases .
Approach: They propose a leaderboard to assess different LLMs' long-context comprehension abilities . they offer open-source access to the benchmarks and maintain a dedicated website .
Outcome: The proposed model assesses different LLMs on selected benchmarks and provides open-source access to the benchmarks.
Data Diversity Matters for Robust Instruction Tuning (2024.findings-emnlp)

Copied to clipboard

Challenge: Recent studies have shown that by curating high quality and diverse instruction tuning datasets, we can significantly improve instruction-following capabilities.
Approach: They propose an algorithm to control diversity and quality of instruction tuning datasets and validate it.
Outcome: The proposed algorithm significantly improves worst and average case performance on large scale instruction tuning datasets.
Enhancing Learning-Based Binary Code Similarity Detection Model through Adversarial Training with Multiple Function Variants (2024.findings-emnlp)

Copied to clipboard

Challenge: Existing Learning-Based Binary Code Similarity Detection (LB-BCSD) methods exhibit lower accuracy in recognizing functions with the same functionality but different implementations.
Approach: They propose a gradient-guided adversarial attack method based on critical code called FuncFooler which perturbs critical code to generate multiple variants of the same function.
Outcome: The proposed method increases the accuracy of the current Learning-Based Binary Code Similarity Detection (LB-BCSD) model by 5%-7%.
Adversarial Preference Learning for Robust LLM Alignment (2025.findings-acl)

Copied to clipboard

Challenge: Modern language models rely on Reinforcement Learning from Human Feedback (RLHF) to encourage safe behaviors, but they remain vulnerable to adversarial attacks due to three key limitations: (1) the inefficiency and high cost of human annotation; (2) the vast diversity of potential adversarials; and (3) the risk of feedback bias and reward hacking.
Approach: They propose an iterative adversarial training method that incorporates three key innovations to address these challenges.
Outcome: Experiments on Mistral-7B-Instruct-v0.3 show that the proposed method significantly enhances robustness and reduces harmful outputs from 5.88% to 0.43%.
SEQZERO: Few-shot Compositional Semantic Parsing with Sequential Prompts and Zero-shot Models (2022.findings-naacl)

Copied to clipboard

Challenge: Recent research shows promising results on combining pretrained language models with canonical utterance for few-shot semantic parsing.
Approach: They propose a few-shot semantic parsing method that decomposes a problem into a sequence of sub-problems, which correspond to the sub-clauses of the formal language.
Outcome: The proposed method achieves SOTA performance of BART-based models on GeoQuery and EcommerceQuery, which are two few-shot datasets with compositional data split.
Disentangled Code Representation Learning for Multiple Programming Languages (2021.findings-acl)

Copied to clipboard

Challenge: Developing effective distributed representations of source code is challenging . current code embedding approaches that represent the semantic and syntax of code are less interpretable .
Approach: They propose a disentangled code representation learning approach to separate the semantic from the syntax of source code under a multi-programming-language setting.
Outcome: The proposed approach achieves better interpretability and generalizability over existing methods.
SUBS: Subtree Substitution for Compositional Semantic Parsing (2022.naacl-main)

Copied to clipboard

Challenge: Semantic parsing models fail at compositional generalization due to lack of reasoning ability.
Approach: They propose to use subtree substitution for compositional data augmentation to increase the number of subtreas with similar semantic functions as exchangeable.
Outcome: The proposed method improves performance on Scan and GeoQuery, and new SOTA on compositional split of GeoQuery.

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