Profiles2020-2026ORCID 0000-0001-7290-0487
Hai Zhao
Shanghai Jiao Tong University
101Papers
60Last author
77Unique co-authors
0First author
0Solo
First and last author come from byline position, which some venues assign alphabetically.
Recent papers101
- Remember Me, Refine Me: A Dynamic Procedural Memory Framework for Experience-Driven Agent EvolutionAuthor 7 of 7ACL Findings 2026 · 2 citations
- RACER: Retrieval-Augmented Contextual Rapid Speculative DecodingAuthor 5 of 5ACL Findings 2026 · 1 citation
- Scaling LLM Speculative Decoding: Non-Autoregressive Forecasting in Large-Batch ScenariosAuthor 6 of 6AAAI 2026 · 1 citation
- Faster MoE LLM Inference for Extremely Large ModelsAuthor 7 of 7ACL Findings 2026
- GoT-R1: Internalizing Graph-of-Thought via Structural Reinforcement for High-Density ReasoningAuthor 4 of 6ACL Findings 2026
- PAR: Training-Free Positional Perturbation and Attention Recycling for Faithful OCRAuthor 5 of 5ACL 2026
- ParaCook: On Time-Efficient Planning for Multi-Agent SystemsAuthor 9 of 10ACL Findings 2026
- Caution for the Environment: Multimodal LLM Agents are Susceptible to Environmental DistractionsAuthor 7 of 7ACL 2025 · 54 citations
- SCANS: Mitigating the Exaggerated Safety for LLMs via Safety-Conscious Activation SteeringAuthor 3 of 3AAAI 2025 · 35 citations
- SmallKV: Small Model Assisted Compensation of KV Cache Compression for Efficient LLM InferenceAuthor 6 of 7NeurIPS 2025 · 8 citations
- DAC: A Dynamic Attention-aware Approach for Task-Agnostic Prompt CompressionAuthor 3 of 5ACL 2025 · 7 citations
- Towards Enhanced Immersion and Agency for LLM-based Interactive DramaAuthor 5 of 5ACL 2025 · 7 citations
- LESA: Learnable LLM Layer Scaling-UpAuthor 7 of 7ACL 2025 · 6 citations
- Wide-Horizon Thinking and Simulation-Based Evaluation for Real-World LLM Planning with Multifaceted ConstraintsAuthor 7 of 7NeurIPS 2025 · 6 citations
- X-TURING: Towards an Enhanced and Efficient Turing Test for Long-Term Dialogue AgentsAuthor 3 of 3ACL 2025 · 6 citations
- KV-Latent: Dimensional-level KV Cache Reduction with Frequency-aware Rotary Positional EmbeddingAuthor 6 of 6ACL 2025 · 5 citations
- IAM: Efficient Inference through Attention Mapping between Different-scale LLMsAuthor 3 of 3ACL 2025 · 3 citations
- MEGen: Generative Backdoor into Large Language Models via Model EditingAuthor 4 of 6ACL Findings 2025 · 3 citations
- CoViPAL: Layer-wise Contextualized Visual Token Pruning for Large Vision-Language ModelsAuthor 6 of 8EMNLP Findings 2025 · 1 citation
- Segment First or Comprehend First? Explore the Limit of Unsupervised Word Segmentation with Large Language ModelsAuthor 5 of 6ACL 2025 · 1 citation
- Can Large Language Models Be Good Language Teachers?Author 5 of 6EMNLP 2025
- Dialogue-RAG: Enhancing Retrieval for LLMs via Node-Linking Utterance RewritingAuthor 6 of 6ACL 2025
- Faster In-Context Learning for LLMs via N-Gram Trie Speculative DecodingAuthor 7 of 8EMNLP 2025
- From Parameters to Performance: A Data-Driven Study on LLM Structure and DevelopmentAuthor 5 of 7EMNLP 2025
- Game Development as Human-LLM InteractionAuthor 3 of 3ACL 2025
- XQuant: Achieving Ultra-Low Bit KV Cache Quantization with Cross-Layer CompressionAuthor 6 of 6EMNLP 2025
- Chinese Spelling Correction as Rephrasing Language ModelAuthor 3 of 3AAAI 2024 · 36 citations
- PyramidInfer: Pyramid KV Cache Compression for High-throughput LLM InferenceAuthor 6 of 6ACL Findings 2024 · 31 citations
- A Novel Energy Based Model Mechanism for Multi-Modal Aspect-Based Sentiment AnalysisAuthor 5 of 5AAAI 2024 · 22 citations
- LaCo: Large Language Model Pruning via Layer CollapseAuthor 3 of 3EMNLP Findings 2024 · 21 citations
Showing the 30 most recent indexed papers.
