Papers by Haibo Wang

26 papers
PromptPrism: A Linguistically-Inspired Taxonomy for Prompts (2026.findings-eacl)

Copied to clipboard

Challenge: PromptPrism is a linguistically-inspired taxonomy that enables prompt analysis across three hierarchical levels.
Approach: They propose a linguistically-inspired taxonomy that enables prompt analysis across three hierarchical levels: functional structure, semantic component, and syntactic pattern.
Outcome: The proposed taxonomy bridges traditional language understanding with modern LLM research . it improves prompt quality and improves model performance across tasks .
Disentangle to Decay: Linear Attention with Trainable Decay Factor (2025.coling-main)

Copied to clipboard

Challenge: Existing linear attention models use a decay factor based positional encoding (PE), but the decay factor is manually designed and non-trainable, limiting further optimization.
Approach: They propose a PE-based positional encoding that disentangles decay factor into two parts to achieve further optimization and stable training.
Outcome: The proposed model achieves stable training of decay factor and improves inference efficiency in normal context and extrapolation scenarios.
SLOT: Structuring the Output of Large Language Models (2025.emnlp-industry)

Copied to clipboard

Challenge: Structured outputs are essential for large language models (LLMs) but often deviate from predefined schemas hampering reliable application development.
Approach: They propose a model-agnostic approach that transforms unstructured LLM outputs into precise structured formats.
Outcome: The proposed model-agnostic approach transforms unstructured LLM outputs into precise structured formats.
A Systematic Survey of Automatic Prompt Optimization Techniques (2025.emnlp-main)

Copied to clipboard

Challenge: Recent advances in prompt engineering have created impediments for end users to adopt . however, prompt engineering remains an impedance due to rapid advances in models, tasks, and associated best practices.
Approach: They propose to define APO as a 5-part unifying framework and categorize all relevant works based on their salient features.
Outcome: The proposed framework aims to improve the performance of large language models on various tasks.
Understanding and Improving the Robustness of Terminology Constraints in Neural Machine Translation (2023.acl-long)

Copied to clipboard

Challenge: Existing terminology constraint test sets are blind to this issue due to oversimplified settings . PH methods retain high constraint accuracy but lower translation quality .
Approach: They propose a method that replaces terminology terms with ordered labels . placeholder methods are better at retaining high constraint accuracy but lower translation quality .
Outcome: The proposed method achieves high accuracy and translation quality regardless of the number or length of constraints.
Why Supervised Fine-Tuning Fails to Learn: A Systematic Study of Incomplete Learning in Large Language Models (2026.acl-long)

Copied to clipboard

Challenge: Incomplete learning is widespread and heterogeneous in large language models . authors identify five recurrent sources of incomplete learning: missing prerequisite knowledge, conflicts between SFT supervision and pre-training knowledge, internal inconsistencies within SFT data, left-side forgetting during sequential fine-tuning, and insufficient optimization for rare or complex patterns.
Approach: They propose a diagnostic-first framework that maps incomplete learning to causes . they identify five recurrent sources of incomplete learning: missing prerequisite knowledge, conflicts between supervision and pre-training knowledge, internal inconsistencies, left-side forgetting during sequential fine-tuning, and insufficient optimization for rare or complex patterns.
Outcome: The proposed framework maps incomplete learning to causes using observable training and inference signals.
IPR: Intelligent Prompt Routing with User-Controlled Quality-Cost Trade-offs (2025.emnlp-industry)

Copied to clipboard

Challenge: Existing systems require users to manually select models or employ rigid routing rules that fail to capture the continuous spectrum of query complexity.
Approach: They propose a quality-constrained intelligent prompt routing framework that automatically selects optimal models based on predicted response quality and user-specified tolerance levels.
Outcome: The proposed framework achieves 43.9% cost reduction while maintaining quality parity with strongest model in the Claude family and processes requests with sub-150ms latency.
Resource-Friendly Dynamic Enhancement Chain for Multi-Hop Question Answering (2025.findings-acl)

Copied to clipboard

Challenge: Existing approaches to solve multi-hop question answering challenges require multiple rounds of retrieval and iterative generation.
Approach: They propose a framework that decomposes complex questions into coherent subquestions . it then iteratively refines these subquests through context-aware rewriting to generate effective query formulations.
Outcome: The proposed framework performs on par with or surpasses state-of-the-art benchmarks while significantly reducing token consumption.
Optimal Transport-Based Token Weighting scheme for Enhanced Preference Optimization (2025.acl-long)

Copied to clipboard

Challenge: Existing methods for direct preference optimization assign equal importance to all tokens while humans focus on more meaningful parts.
Approach: They propose to use a transport-based token weighting scheme to enhance direct preference optimization by emphasizing meaningful token pairs and de-emphasizing less relevant ones to yield a more contrastive reward difference estimate.
Outcome: Extensive experiments have validated the proposed method in improving instruction-following ability across various settings.
Reason Only When Needed: Efficient Generative Reward Modeling via Model-Internal Uncertainty (2026.findings-acl)

Copied to clipboard

Challenge: Existing approaches to generating reward models rely on voting-based mechanisms to evaluate CoT outputs.
Approach: They propose an efficient generative reward modeling framework grounded in model-internal uncertainty.
Outcome: The proposed framework reduces inference cost while improving answer accuracy.
“Yes, My LoRD.” Guiding Language Model Extraction with Locality Reinforced Distillation (2025.acl-long)

Copied to clipboard

Challenge: Existing methods for model extraction attacks on large language models are inadequate . existing methods neglect the inconsistency between training tasks and LLM alignment .
Approach: They propose a model extraction algorithm that uses a policy-gradient-style training task to guide the crafting of preference for the local model.
Outcome: The proposed algorithm reduces query complexity while mitigating watermark protection . it can extract various state-of-the-art commercial LLMs while minimizing query complexity .
Diffusion Language Model Inference with Monte Carlo Tree Search (2026.findings-eacl)

Copied to clipboard

Challenge: Existing methods for inference use heuristics to determine which positions to unmask and which tokens to commit . MEDAL is an inference-time scaling framework that integrates Monte Carlo Tree SEarch initialization for Diffusion Language Model inference.
Approach: They propose a framework that integrates Monte Carlo Tree SEarch initialization for Diffusion Language Model inference.
Outcome: The proposed framework achieves 22.0% improvement over existing inference strategies across multiple benchmarks.
Black-Box Visual Prompt Engineering for Mitigating Object Hallucination in Large Vision Language Models (2025.naacl-short)

Copied to clipboard

Challenge: Existing methods to mitigate object hallucination are impractical for proprietary LVLMs.
Approach: They propose a framework to identify optimal visual prompts that enhance LVLM responses without access to model internals.
Outcome: The proposed approach is model-agnostic and can be used on open-source and proprietary LVLMs.
AudioStealer: Extracting Audio Prompts via Shapley Value-Guided Query Search (2026.findings-acl)

Copied to clipboard

Challenge: prompt stealing is a new form of attack that aims to reconstruct high-value prompts that guide music generation.
Approach: They propose a method to steal music prompts from audio domains using a black-box attack framework.
Outcome: The proposed method recovers prompts with high textual consistency to the ground truth while maintaining strong perceptual similarity to the target recordings.
GLIMPSE: Do Large Vision-Language Models Truly Think With Videos or Just Glimpse at Them? (2025.emnlp-main)

Copied to clipboard

Challenge: Existing video benchmarks often resemble image-based questions with scans of only a few key frames, without deep temporal reasoning.
Approach: They propose a video benchmark to assess whether large vision-language models can genuinely think with videos rather than perform superficial frame-level analysis.
Outcome: The proposed benchmark consists of 3,269 videos and over 4,342 highly visual-centric questions across 11 categories, including Trajectory Analysis, Temporal Reasoning, and Forensics Detection.
Crafting Adversarial Inputs for Large Vision-Language Models Using Black-Box Optimization (2026.findings-eacl)

Copied to clipboard

Challenge: Existing white-box jailbreak methods require full model accessibility and require computational costs.
Approach: They propose a black-box jailbreak attack using Zeroth-Order optimization using ZO-SPSA.
Outcome: The proposed method achieves highest jailbreak success rate on three LVLMs, including InstructBLIP, LLaVA and MiniGPT-4.
Retrieval as Attention: End-to-end Learning of Retrieval and Reading within a Single Transformer (2022.emnlp-main)

Copied to clipboard

Challenge: eschewing separate architecture and training for knowledge-intensive tasks is cumbersome . end-to-end training only based on supervision from the end task is awkward .
Approach: They propose a single Transformer that performs retrieval as attention and end-to-end training solely based on supervision from the end QA task.
Outcome: The proposed model outperforms state-of-the-art retrievers and readers on in-domain datasets.
Exploring the Vulnerability of the Content Moderation Guardrail in Large Language Models via Intent Manipulation (2025.findings-emnlp)

Copied to clipboard

Challenge: Prior work has shown that intent detection enhances LLMs’ moderation guardrails, but the robustness of these guardrail mechanisms under malicious manipulations remains under-explored.
Approach: They propose a two-stage intent-based prompt-refinement framework that first transforms harmful inquiries into structured outlines and further reframes them into declarative-style narratives.
Outcome: The proposed framework outperforms several cutting-edge jailbreak methods and evades even advanced Intent Analysis (IA) and Chain-of-Thought (CoT)-based defenses.
Correct after Answer: Enhancing Multi-Span Question Answering with Post-Processing Method (2024.findings-emnlp)

Copied to clipboard

Challenge: Prior work focuses on designing specific methods or applying heuristic strategies to encourage models to predict more correct predictions.
Approach: They propose a framework that uses a post-processing strategy to handle incorrect predictions.
Outcome: The proposed framework significantly improves the Exact Match scores on multiple MSQA datasets.
CreditDecoding: Accelerating Parallel Decoding in Diffusion Large Language Models with Trace Credit (2026.acl-long)

Copied to clipboard

Challenge: Diffusion large language models generate text through iterative denoising with bidirectional attention, enabling richer contextual dependencies.
Approach: They propose a training-free parallel decoding method that fuses Trace Credit with current logits to boost the confidence of correct but underconfident tokens.
Outcome: The proposed method achieves 5.48 times speedup with +0.48 accuracy on LLaDA-8B and is orthogonal to mainstream inference optimizations.
Learning to Conceal Risk: Controllable Multi-turn Red Teaming for LLMs in the Financial Domain (2026.acl-long)

Copied to clipboard

Challenge: Large Language Models (LLMs) are increasingly deployed in high-stakes domains such as finance where unsafe behavior can lead to serious regulatory risks.
Approach: They propose a black-box multi-turn risk-concealed redteaming framework that progressively conceals surface-level risk while exploiting regulatory-violating behaviors.
Outcome: Experiments on nine widely used LLMs show that the proposed framework achieves 93.19% average attack success rate (ASR) and improves the average ASR to 95.00%.
Grounded-VideoLLM: Sharpening Fine-grained Temporal Grounding in Video Large Language Models (2025.findings-emnlp)

Copied to clipboard

Challenge: Video Large Language Models (VLMs) have been praised for their performance in coarse-grained video understanding but still face ineffective temporal grounding and inadequate timestamp representations.
Approach: They propose a novel Video-LLM that senses and reasoned over specific video moments with fine-grained temporal precision.
Outcome: The proposed model surpasses existing models in fine-grained video understanding tasks and exhibits strong potential as a general video understanding assistant.
KnowTuning: Knowledge-aware Fine-tuning for Large Language Models (2024.emnlp-main)

Copied to clipboard

Challenge: Large language models (LLMs) are a default solution for many natural language processing tasks.
Approach: They propose a knowledge-aware fine-tuning method to improve LLMs' knowledge awareness . they propose augmentation and comparison stages to improve accuracy and reliability .
Outcome: The proposed method generates more facts with less factual error rate under fine-grained facts evaluation.
GOVERN: Gradient Orientation Vote Ensemble for Multi-Teacher Reinforced Distillation (2024.emnlp-industry)

Copied to clipboard

Challenge: Pre-trained language models have achieved remarkable performance in OpenQA, but for practical deployment, knowledge distillation is crucial to maintain high performance while operating under computational constraints.
Approach: They propose an algorithm to perform unsupervised knowledge distillation without the guidance of labels to achieve 99.5% of performance.
Outcome: The proposed algorithm achieves 99.5% of performance in a commercial question-answering system.
Now You Hear Me: Audio Narrative Attacks Against Large Audio–Language Models (2026.eacl-long)

Copied to clipboard

Challenge: Existing jailbreaks against large audio-language models fall into two categories . early work converted text-based prompts into synthetic speech, while subsequent work introduced minor acoustic variations such as accent shifts, phonetic spellings, or stress patterns.
Approach: They propose a text-to-audio jailbreak that embeds disallowed directives within a narrative-style audio stream.
Outcome: The proposed attack exploits structural and acoustic properties of a text-to-audio model . it achieves 98.26% success rate, significantly exceeding baselines for text-based models .
Orchestrating Tokens and Sequences: Dynamic Hybrid Policy Optimization for RLVR (2026.findings-acl)

Copied to clipboard

Challenge: Existing RLVR algorithms focus on different granularities and have complementary strengths and limitations.
Approach: They propose a framework for reinforcement learning with verifiable rewards that bridges RLVR and GSPO . group-level importance ratios are used to update a policy, which preserves fine-grained credit assignment .
Outcome: The proposed framework outperforms existing methods on seven reasoning benchmarks.

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