Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileWorks210 from public data
- Math-Shepherd: Verify and Reinforce LLMs Step-by-step without Human Annotations1,249
An innovative process-oriented math process reward model called Math-Shepherd, which assigns a reward score to each step of math problem solutions, which holds significant potential for the future evolution of LLMs.
- A Survey on In-context Learning1,205
This paper presents a formal definition of ICL and clarify its correlation to related studies, and organizes and discusses advanced techniques, including training strategies, prompt designing strategies, and related analysis.
- Knowledge Neurons in Pretrained Transformers785
This paper examines the fill-in-the-blank cloze task for BERT and proposes a knowledge attribution method to identify the neurons that express the fact, finding that the activation of such knowledge neurons is positively correlated to the expression of their corresponding facts.
- Why Can GPT Learn In-Context? Language Models Implicitly Perform Gradient Descent as Meta-Optimizers420
Inspired by the understanding of meta-optimization, a momentum-based attention is designed by analogy with the momentum- based gradient descent algorithm and its consistently better performance over vanilla attention supports the understanding from another aspect.
- Table-to-Text Generation by Structure-Aware Seq2seq Learning286
The attention visualizations and case studies show that the novel structure-aware seq2seq architecture which consists of field-gating encoder and description generator with dual attention is capable of generating coherent and informative descriptions based on the comprehensive understanding of both the content and the structure of a table.
- 277
- CBLUE: A Chinese Biomedical Language Understanding Evaluation Benchmark262
The first Chinese Biomedical Language Understanding Evaluation Evaluation (CBLUE) benchmark is presented: a collection of natural language understanding tasks including named entity recognition, information extraction, clinical diagnosis normalization, single-sentence/sentence-pair classification, and an associated online platform for model evaluation, comparison, and analysis.
- Jointly Extracting Event Triggers and Arguments by Dependency-Bridge RNN and Tensor-Based Argument Interaction261
A novel dependency bridge recurrent neural network (dbRNN) is proposed that simultaneously applying tree structure and sequence structure in RNN brings much better performance than only uses sequential RNN.
- Speculative Decoding: Exploiting Speculative Execution for Accelerating Seq2seq Generation217
Experimental results on various seq2seq tasks including machine translation and abstractive summarization show the proposed Speculative Decoding approach can achieve around $5\times$ speedup for the popular Transformer architectures with comparable generation quality to beam search decoding, refreshing the impression that the draft-then-verify paradigm introduces only $1.4\times$$\sim $ speedup.
- A Token-level Reference-free Hallucination Detection Benchmark for Free-form Text Generation210
This work proposes a novel token-level, reference-free hallucination detection task and an associated annotated dataset named HaDeS (HAllucination DEtection dataSet), and utilizes an iterative model-in-loop strategy to mitigate label imbalance during annotation.
- 190
- Towards Time-Aware Knowledge Graph Completion184
A novel time-aware knowledge graph completion model that is able to predict links in a KG using both the existing facts and the temporal information of the facts and achieves the state-of-the-art on temporal facts consistently.
- A Soft-label Method for Noise-tolerant Distantly Supervised Relation Extraction150
This work introduces an entity-pair level denoise method which exploits semantic information from correctly labeled entity pairs to correct wrong labels dynamically during training, and proposes a joint score function which combines the relational scores based on the entity- Pair representation and the confidence of the hard label to obtain a new label.
- 147
- Exploiting collaborative filtering techniques for automatic assessment of student free-text responses144
The experimental results show that the feature-based matrix factorization model outperforms the baseline models and the model with a re-ranking phase can achieve a better and competitive performance -- 63.6% overall accuracy on the Beetle dataset.
- Encoding Temporal Information for Time-Aware Link Prediction136
A novel time-aware KB embedding approach taking advantage of the happening time of facts is proposed, which uses temporal order constraints to model transformation between time-sensitive relations and enforce the embeddings to be temporally consistent and more accurate.
- Synthetic Data (Almost) from Scratch: Generalized Instruction Tuning for Language Models123
This work introduces Generalized Instruction Tuning (called GLAN), a general and scalable method for instruction tuning of Large Language Models (LLMs) that exclusively utilizes a pre-curated taxonomy of human knowledge and capabilities as input and generates large-scale synthetic instruction data across all disciplines.
- 117
- Order-Planning Neural Text Generation From Structured Data117
This paper proposes an order-planning text generation model, where order information is explicitly captured by link-based attention and a self-adaptive gate combines the link- based attention with traditional content-based Attention.
- Implicit Discourse Relation Classification via Multi-Task Neural Networks117
This work designs related discourse classification tasks specific to a corpus, and proposes a novel Convolutional Neural Network embedded multi-task learning system to synthesize these tasks by learning both unique and shared representations for each task.
- Incorporating Glosses into Neural Word Sense Disambiguation107
GAS models the semantic relationship between the context and the gloss in an improved memory network framework, which breaks the barriers of the previous supervised methods and knowledge-based methods.
- Making Large Language Models Better Reasoners with Alignment94
The constraint alignment loss has two objectives: a) Alignment, which guarantees that positive scores surpass negative scores to encourage answers with high-quality COTs, and b) Constraint, which keeps the negative scores confined to a reasonable range to prevent the model degradation.
- XGPT: Cross-modal Generative Pre-Training for Image Captioning86
XGPT is a new method of Cross-modal Generative Pre-Training for Image Captioning that is designed to pre-train text-to-image caption generators through three novel generation tasks, including Image-conditioned Masked Language Modeling (IMLM), Image- Conditioned Denoising Autoencoding (IDA), and Text-conditioning Image Feature Generation (TIFG).
- Stabilizing MoE Reinforcement Learning by Aligning Training and Inference Routers73
Rollout Routing Replay (R3), a method that records routing distributions from the inference engine and replays them during training significantly reduces training-inference policy KL divergence and mitigates extreme discrepancies without compromising training speed, is proposed.
- An Enhanced Span-based Decomposition Method for Few-Shot Sequence Labeling73
SD decomposes the span matching problem into a series of span-level procedures, mainly including enhanced span representation, class prototype aggregation and span conflicts resolution, which achieves the new state-of-the-art results on two popular FSSL benchmarks, FewNERD and SNIPS.
Publication data from OpenAlex, with missing venues and authors filled in from Crossref; citation counts are the higher of OpenAlex and Semantic Scholar, last synced 2026-10-15. One-sentence summaries under some papers are written by Semantic Scholar’s model. Citation counts may be lower than on Google Scholar, which indexes more sources.
Report an error
Wrong papers, two people merged into one, or a profile that should not be here? Tell us and we will fix or hide it.