Papers by Ang Lv
Masked Thought: Simply Masking Partial Reasoning Steps Can Improve Mathematical Reasoning Learning of Language Models (2024.acl-long)
Copied to clipboard
Changyu Chen, Xiting Wang, Ting-En Lin, Ang Lv, Yuchuan Wu, Xin Gao, Ji-Rong Wen, Rui Yan, Yongbin Li
| Challenge: | Despite the advances in large language models, they still face difficulties with multi-step reasoning tasks. |
| Approach: | They propose a method that randomly masks certain tokens within the chain of thought to improve model accuracy by 5% over standard supervised fine-tuning. |
| Outcome: | The proposed method improves accuracy and accuracy by 5% over standard fine-tuning with a few codes modified. |
Union-of-Experts: Neurons in Mixture-of-Experts are Secretly Routers (2026.acl-long)
Copied to clipboard
| Challenge: | Mixture-of-Experts (MoE) models rely on an external router to assign tokens to experts, resulting in suboptimal performance. |
| Approach: | They propose an MoE variant that performs "expert-autonomous routing" by pre-designating a fraction of neurons within each expert as "routing neurons" they pre-train UoE models with up to 3B parameters and show they outperform traditional MoEs with matched efficiency. |
| Outcome: | The proposed model outperforms existing models with 3B parameters and provides valuable insights into expert-autonomous selection and the broader routing mechanisms of MoE models. |
DialoGPS: Dialogue Path Sampling in Continuous Semantic Space for Data Augmentation in Multi-Turn Conversations (2023.acl-long)
Copied to clipboard
| Challenge: | a context leads to various responses, and a response answers multiple contexts. |
| Approach: | They propose a method that augments open-domain dialogue generation from a many-to-many perspective. |
| Outcome: | The proposed method can augment open-domain dialogue generation tasks with automatic and human evaluation. |
Data Pollination: An Emergent Ecological Process Driving AI Population Evolution (2026.acl-long)
Copied to clipboard
| Challenge: | evidence from deployed systems suggests that language models interact through a shared data ecosystem. |
| Approach: | They propose to use data pollination to investigate stability dynamics under synthetic data training to investigate model collapse. |
| Outcome: | The proposed model can mitigate model collapse observed in recursive training, and improve performance across benchmarks. |
Mixture-of-Modules: Reinventing Transformers as Dynamic Assemblies of Modules (2024.emnlp-main)
Copied to clipboard
| Challenge: | Empirical results show that MoMs consistently outperform vanilla transformers . |
| Approach: | They propose an architecture that allows for a mixture-of-modules computation that uses a finite set of modules defined by multi-head attention and feed-forward networks. |
| Outcome: | The proposed architecture outperforms vanilla Transformers and their variants in multiple ways. |
Language Models “Grok” to Copy (2025.naacl-short)
Copied to clipboard
| Challenge: | We examine the pre-training dynamics of language models, focusing on their ability to copy text from preceding context. |
| Approach: | They propose that Transformer-based language models develop copying abilities similarly to grokking . they argue that the connection between groking and context copying can improve in-context performance. |
| Outcome: | The proposed model development is similar to grokking, but the speed is independent of tokens trained. |
StepHint: Multi-level Stepwise Hints Enhance Reinforcement Learning to Reason (2026.acl-long)
Copied to clipboard
| Challenge: | Reinforcement learning with verifiable rewards (RLVR) approaches face two challenges: the near-miss reward problem and exploration stagnation. |
| Approach: | They propose an algorithm that partitions valid reasoning chains into reasoning steps using multi-level stepwise hints. |
| Outcome: | The proposed method outperforms competing RLVR enhancement methods across six mathematical benchmarks and two out-of-domain benchmarks. |
Fortify the Shortest Stave in Attention: Enhancing Context Awareness of Large Language Models for Effective Tool Use (2024.acl-long)
Copied to clipboard
| Challenge: | In this paper, we demonstrate that an inherent waveform pattern in the attention allocation of large language models significantly affects their performance in tasks demanding a high degree of context awareness. |
| Approach: | They propose a method that compensates an attention trough with an attention peak by a process to enhance the model's awareness to various contextual positions. |
| Outcome: | The proposed method improves the performance of a 7B model on the largest tool-use benchmark, comparable to that of GPT-4. |
An Analysis and Mitigation of the Reversal Curse (2024.emnlp-main)
Copied to clipboard
| Challenge: | Recent research observes a phenomenon in large language models called the "reversal curse" when dealing with two entities, LLMs excel in handling sequences in the form of "aRb" but when asked "who is Mary Lee Pfeiffer's son?" the LLM exhibits considerable confusion and fails to provide a as the answer . |
| Approach: | They conduct the first-ever study of how the reversal curse happens in large language models . they find that LLMs excel in handling sequences in the form of "aRb" but struggle to provide a satisfactory answer when asked "who is Mary Lee Pfeiffer's son?" |
| Outcome: | The proposed study shows that the reversal curse can stem from specific training objectives . the study also shows that a reverse query can be difficult to understand . |
Envisioning Future from the Past: Hierarchical Duality Learning for Multi-Turn Dialogue Generation (2023.acl-long)
Copied to clipboard
| Challenge: | Existing approaches to generate high quality responses rely on future text . |
| Approach: | They propose a hierarchical duality learning for dialogue to simulate human cognitive ability . they utilize hierarchically dualities at token hierarchy and utterance hierarchy to simulate duality . |
| Outcome: | The proposed model can generate high quality responses that connect both previous and follow-up dialogues. |
HoPE: A Novel Positional Encoding Without Long-Term Decay for Enhanced Context Awareness and Extrapolation (2025.acl-long)
Copied to clipboard
| Challenge: | Existing positional encodings exhibit long-term decay, based on an entrenched and long-standing opinion that tokens farther away from the current position carry less relevant information. |
| Approach: | They propose a high-frequency rotary position encoding (HoPE) that replaces specific components in RoPE with position-independent ones, retaining only high- frequency signals. |
| Outcome: | The proposed method exhibits greater robustness to the out-of-distribution behavior in attention patterns during extrapolation. |
More is not always better? Enhancing Many-Shot In-Context Learning with Differentiated and Reweighting Objectives (2025.acl-long)
Copied to clipboard
Xiaoqing Zhang, Ang Lv, Yuhan Liu, Flood Sung, Wei Liu, Jian Luan, Shuo Shang, Xiuying Chen, Rui Yan
| Challenge: | Large language models excel at few-shot in-context learning but performance plateaus as ICL demonstrations increase from a few to many. |
| Approach: | They propose a novel optimization method that optimizes the negative log-likelihood objective and reweights the model to achieve many-shot performance. |
| Outcome: | The proposed method achieves significant performance improvements across a large-scale dataset. |
Batch-ICL: Effective, Efficient, and Order-Agnostic In-Context Learning (2024.findings-acl)
Copied to clipboard
| Challenge: | In-context learning (ICL) is a meta-optimization process that affects performance . we develop a batch-based inference algorithm that is order-agnostic to ICL examples . |
| Approach: | They develop an order-agnostic inference algorithm that aggregates ICL examples in batches . they find it outperforms most permutations of ICL, and it even exceeds the best order . |
| Outcome: | The proposed method outperforms standard ICL examples while reducing computational resources. |