Papers by Boyuan Li
LLM-Powered Benchmark Factory: Reliable, Generic, and Efficient (2026.acl-long)
Copied to clipboard
Peiwen Yuan, Shaoxiong Feng, Yiwei Li, Xinglin Wang, Yueqi Zhang, Jiayi Shi, Chuyi Tan, Boyuan Pan, Yao Hu, Kan Li
| Challenge: | Using generic and efficient benchmark generators, human annotators are limited by inefficiency . current benchmark generator methods rely on seed signals, leading to long cycles and high costs . |
| Approach: | They propose a framework to evaluate LLMs as generic benchmark generators and integrate them as BenchMaker. |
| Outcome: | The proposed framework achieves comparable performance to human-annotated benchmarks on most metrics. |
Make Every Penny Count: Difficulty-Adaptive Self-Consistency for Cost-Efficient Reasoning (2025.findings-naacl)
Copied to clipboard
Xinglin Wang, Shaoxiong Feng, Yiwei Li, Peiwen Yuan, Yueqi Zhang, Chuyi Tan, Boyuan Pan, Yao Hu, Kan Li
| Challenge: | Existing decoding strategies for chain-of-thought reasoning do not exploit prior information about question difficulty. |
| Approach: | They propose a decoding strategy called self-consistency to improve reasoning performance by adjusting the number of samples based on the posterior distribution of a set of pre-samples. |
| Outcome: | The proposed method outperforms baseline methods on arithmetic, commonsense and symbolic reasoning tasks while achieving comparable performance. |
Focused Large Language Models are Stable Many-Shot Learners (2024.emnlp-main)
Copied to clipboard
Peiwen Yuan, Shaoxiong Feng, Yiwei Li, Xinglin Wang, Yueqi Zhang, Chuyi Tan, Boyuan Pan, Heda Wang, Yao Hu, Kan Li
| Challenge: | In-Context Learning (ICL) enables large language models to achieve rapid task adaptation by learning from demonstrations. |
| Approach: | They propose a training-free method that disperses model attention from the query . they propose 'focus' search strategy that uses model perplexity to ensure sufficient attention . |
| Outcome: | The proposed method achieves an average performance improvement of 5.2% over vanilla ICL and scales well with many-shot demonstrations. |
InsBank: Evolving Instruction Subset for Ongoing Alignment (2025.findings-emnlp)
Copied to clipboard
Jiayi Shi, Yiwei Li, Shaoxiong Feng, Peiwen Yuan, Xinglin Wang, Yueqi Zhang, Chuyi Tan, Boyuan Pan, Huan Ren, Yao Hu, Kan Li
| Challenge: | Recent studies emphasize that quality and diversity of instruction data are more crucial than quantity, highlighting the need to select diverse, high-quality subsets to reduce training costs. |
| Approach: | They propose to use a continuously updated repository to integrate the latest valuable instruction data with a progressive evolution framework to evolve InsBank over time. |
| Outcome: | The proposed framework outperforms baselines in InsBank evolution and extracts budget-specific subsets. |
Integrate the Essence and Eliminate the Dross: Fine-Grained Self-Consistency for Free-Form Language Generation (2024.acl-long)
Copied to clipboard
| Challenge: | Existing methods to improve output quality without aggregating input tokens are limited by the complexity of aggregation of responses. |
| Approach: | They propose to extract and integrate segment-level commonalities from candidate samples to enhance performance of LLMs in open-ended and reasoning tasks. |
| Outcome: | The proposed method improves performance on reasoning, code generation and mathematical reasoning tasks without requiring additional models and overlooking the knowledge present among the candidates. |
Beyond One-Size-Fits-All: Tailored Benchmarks for Efficient Evaluation (2025.acl-long)
Copied to clipboard
Peiwen Yuan, Yueqi Zhang, Shaoxiong Feng, Yiwei Li, Xinglin Wang, Jiayi Shi, Chuyi Tan, Boyuan Pan, Yao Hu, Kan Li
| Challenge: | Existing efficient methods estimate performance of models on large benchmarks, but these methods rely on the assumption that target models have high prediction consistency with source models. |
| Approach: | They propose a method that conducts customized evaluation tailored to each target model. |
| Outcome: | The proposed method reduces the MAE of estimates by 31.4% on benchmarks across 300 models. |
SafeMT: Multi-turn Safety for Multimodal Language Models (2026.acl-long)
Copied to clipboard
Han Zhu, Juntao Dai, Jiaming Ji, Haoran Li, Chengkun Cai, Pengcheng Wen, Chi-Min Chan, Boyuan Chen, Yaodong Yang, Sirui Han, Yike Guo
| Challenge: | Multi-turn dialogues pose a greater risk than single prompts, but existing safety benchmarks do not account for this situation. |
| Approach: | They propose a benchmark that features dialogues of varying lengths generated from harmful queries accompanied by images. |
| Outcome: | The proposed model reduces multi-turn Attack Success Rate (ASR) compared to existing guard models. |
Learn To Remember: Transformer with Recurrent Memory for Document-Level Machine Translation (2022.findings-naacl)
Copied to clipboard
| Challenge: | Recent studies have shown that the effective use of contextual information between sentences can achieve better performance in document-level machine translation. |
| Approach: | They propose a recurrent memory unit to the Transformer to support the information exchange between the sentence and previous context. |
| Outcome: | The proposed model outperforms the previous work on TED and News by 0.91 s-BLEU and 1.49 d-BLUE on average. |
IGenBench: Benchmarking the Reliability of Text-to-Infographic Generation (2026.acl-long)
Copied to clipboard
Yinghao Tang, Xueding Liu, Boyuan Zhang, Tingfeng Lan, Yupeng Xie, Jiale Lao, Yiyao Wang, Haoxuan Li, Tingting Gao, Bo Pan, Luoxuan Weng, Xiuqi Huang, Minfeng Zhu, Yingchaojie Feng, Yuyu Luo, Wei Chen
| Challenge: | Generated infographics may appear correct at first glance but contain easily overlooked issues, such as distorted data encoding or incorrect textual content. |
| Approach: | They propose to evaluate reliability of text-to-infographic generation using IGenBench . they employ multimodal large language models to verify each question . |
| Outcome: | The proposed framework decomposes reliability verification into atomic yes/no questions based on a taxonomy of 10 question types. |
CogLM: Tracking Cognitive Development of Large Language Models (2025.naacl-long)
Copied to clipboard
| Challenge: | Large Language Models (LLMs) have recently shown remarkable abilities across a wide variety of tasks, but few studies have explored the reasons behind the evolutionary relationship among various abilities. |
| Approach: | They construct a benchmark CogLM based on Piaget's Theory of Cognitive Development (PTC) which measures the cognitive levels of Large Language Models (LLMs) using 1,220 questions spanning 10 cognitive abilities crafted by more than 20 human experts. |
| Outcome: | The proposed framework provides a comprehensive testbed for the cognitive levels of LLMs. |
Speculative Decoding for Multi-Sample Inference (2025.findings-emnlp)
Copied to clipboard
Yiwei Li, Jiayi Shi, Shaoxiong Feng, Peiwen Yuan, Xinglin Wang, Yueqi Zhang, Ji Zhang, Chuyi Tan, Boyuan Pan, Yao Hu, Kan Li
| Challenge: | Speculative decoding method exploits consensus of parallel reasoning paths to synthesize high-quality draft tokens without auxiliary models or external databases. |
| Approach: | They propose a speculative decoding method that exploits the consensus of parallel reasoning paths to synthesize high-quality draft tokens without auxiliary models or external databases. |
| Outcome: | The proposed method exploits the intrinsic consensus of parallel reasoning paths to synthesize high-quality draft tokens without auxiliary models or databases. |
PKU-SafeRLHF: Towards Multi-Level Safety Alignment for LLMs with Human Preference (2025.acl-long)
Copied to clipboard
Jiaming Ji, Donghai Hong, Borong Zhang, Boyuan Chen, Josef Dai, Boren Zheng, Tianyi Alex Qiu, Jiayi Zhou, Kaile Wang, Boxun Li, Sirui Han, Yike Guo, Yaodong Yang
| Challenge: | Using large-scale annotation data, large language models can generate noise, errors and biases, leading to unexpected behaviours. |
| Approach: | They propose a dataset to promote safety alignment in large language models . they separate helpfulness and harmlessness annotations for question-answering pairs . |
| Outcome: | The proposed dataset provides 44.6k prompts and 265k question-answer pairs with safety meta-labels for 19 harm categories and three severity levels, with answers generated by Llama-family models. |
From Sub-Ability Diagnosis to Human-Aligned Generation: Bridging the Gap for Text Length Control via MarkerGen (2025.acl-long)
Copied to clipboard
Peiwen Yuan, Chuyi Tan, Shaoxiong Feng, Yiwei Li, Xinglin Wang, Yueqi Zhang, Jiayi Shi, Boyuan Pan, Yao Hu, Kan Li
| Challenge: | Existing methods to control text length are lacking in LCTG, posing a major limitation for practical applications. |
| Approach: | They propose a plug-and-play approach that decomposes LCTG sub-abilities with human patterns as reference and performs detailed error analysis. |
| Outcome: | The proposed method significantly improves LCTG across various settings, exhibiting outstanding effectiveness and generalizability. |
Revisiting Self-Consistency from Dynamic Distributional Alignment Perspective on Answer Aggregation (2025.findings-acl)
Copied to clipboard
Yiwei Li, Ji Zhang, Shaoxiong Feng, Peiwen Yuan, Xinglin Wang, Jiayi Shi, Yueqi Zhang, Chuyi Tan, Boyuan Pan, Yao Hu, Kan Li
| Challenge: | Existing studies on self-consistency show that it improves reasoning abilities by aggregating diverse stochastic samples. |
| Approach: | They propose a confidence-driven mechanism that dynamically calibrates temperature to align with high probability modes. |
| Outcome: | The proposed method outperforms fixed-diversity baselines on reasoning tasks and improves both average and best-case performance. |
Language Models Resist Alignment: Evidence From Data Compression (2025.acl-long)
Copied to clipboard
Jiaming Ji, Kaile Wang, Tianyi Alex Qiu, Boyuan Chen, Jiayi Zhou, Changye Li, Hantao Lou, Josef Dai, Yunhuai Liu, Yaodong Yang
| Challenge: | Large language models (LLMs) may exhibit undesirable behaviors due to the inevitable biases and harmful content present in training. |
| Approach: | They propose to investigate the elasticity of large language models by examining their performance. |
| Outcome: | The proposed model performance declines rapidly before reverting to the pre-training distribution, the authors show . the proposed model weight and code are available at pku-lm-res ist-alignment.github.io. |
Reinforced Dynamic Reasoning for Conversational Question Generation (P19-1)
Copied to clipboard
| Challenge: | Empirical results on the recently released CoQA dataset demonstrate the effectiveness of our method . large-scale highquality conversational question answering datasets such as CoQA and QuAC can help train models to answer sequential questions. |
| Approach: | They propose a task called Conversational Question Generation which generates a question based on a passage and a conversation history to generate the next question. |
| Outcome: | The proposed method is based on a question-answering style conversation dataset . it can be used to generate meaningful questions on QA and SQuAD datasets . |
Look and Think: Efficient Multimodal Reasoning via Modality-Decoupled Compression (2026.findings-acl)
Copied to clipboard
| Challenge: | Multimodal large language models have strong performance on visual question answering benchmarks . however, their inference efficiency is severely constrained by the rapidly growing context . |
| Approach: | They propose a modality-decoupled compression method that enables efficient multimodal inference . they propose to evict visual tokens whenever visual grounding is unnecessary . |
| Outcome: | The proposed method reduces the average context length by up to 57% while maintaining comparable performance to the standard MLLM baseline. |
Generative Dense Retrieval: Memory Can Be a Burden (2024.eacl-long)
Copied to clipboard
| Challenge: | Empirical results show that Generative Dense Retrieval (GDR) achieves an average of 3.0 R@100 improvement on NQ dataset under multiple settings and has better scalability. |
| Approach: | They propose a Generative Dense Retrieval paradigm that auto-decodes document identifiers given a query and uses memory to avoid memory confusion. |
| Outcome: | Empirical results show that the proposed paradigm improves on the small-scale corpora and improves scalability. |
Poor-Supervised Evaluation for SuperLLM via Mutual Consistency (2024.findings-acl)
Copied to clipboard
| Challenge: | evaluating superLLMs is especially difficult because of their intelligence-intensive nature. |
| Approach: | They propose an evaluation benchmark with accurate labels for SuperLLMs whose capabilities surpass those of humans . they first prove that consistency between model under evaluation and reference model can equalize the true capabilities of the model to be evaluated . |
| Outcome: | The proposed evaluation benchmarks can assess the true capabilities of the model to be evaluated without accurate labels. |
T3: Tree-Autoencoder Constrained Adversarial Text Generation for Targeted Attack (2020.emnlp-main)
Copied to clipboard
| Challenge: | Existing adversarial examples can induce arbitrary errors to the target models, but they can be exploited to estimate robustness of NLP models. |
| Approach: | They propose a target-controllable adversarial attack framework T3 to handle adversarials . they use tree-based decoders to regularize the syntactic correctness of generated text . |
| Outcome: | The proposed framework can be used to estimate the robustness of NLP models. |