14 citations · 31 across the 27 of their papers we have counts for
30 papers
Intern-S2-Preview: Scientific Agentic Foundation Model
Lei Bai, Jiaqi Cao, Chiyu Chen +121
Scientific discovery increasingly requires AI systems that can reason over scientific evidence of heterogeneous modalities, interact with scientific tools and environments, and sus…
Scalable Visual Pretraining for Language Intelligence
Yiming Zhang, Zhonghan Zhao, Wenwei Zhang +14
The rapid progress of large foundation models has been driven predominantly by pretraining on large-scale text corpora. However, many forms of knowledge are conveyed through visual…
Synthetic Pre-Pre-Training Improves Language Model Robustness to Noisy Pre-Training Data
Xu Guo, Runyu Peng, Jian Tong +4
Large language models (LLMs) rely on web-scale corpora for pre-training. The noise inherent in these datasets tends to obscure meaningful patterns and ultimately degrade model perf…
Mousse: Rectifying the Geometry of Muon with Curvature-Aware Preconditioning
Yechen Zhang, Shuhao Xing, Junhao Huang +5
Recent advances in spectral optimization, notably Muon, have demonstrated that constraining update steps to the Stiefel manifold can significantly accelerate training and improve g…
What Makes Position Zero Special? A Mechanistic Study of Position Zero Attention Sinks in LLMs
Runyu Peng, Ruixiao Li, Mingshu Chen +5
Transformers frequently allocate disproportionate attention to specific tokens, a phenomenon known as attention sinks. Causal large language models reliably form one at position ze…
Explicit Multi-head Attention for Inter-head Interaction in Large Language Models
Runyu Peng, Yunhua Zhou, Demin Song +4
In large language models built upon the Transformer architecture, recent studies have shown that inter-head interaction can enhance attention performance. Motivated by this, we pro…
How to Set the Batch Size for Large-Scale Pre-training?
Yunhua Zhou, Junhao Huang, Shuhao Xing +4
The concept of Critical Batch Size, as pioneered by OpenAI, has long served as a foundational principle for large-scale pre-training. However, with the paradigm shift towards the W…
How to Set the Learning Rate for Large-Scale Pre-training?
Yunhua Zhou, Shuhao Xing, Junhao Huang +2
Optimal configuration of the learning rate (LR) is a fundamental yet formidable challenge in large-scale pre-training. Given the stringent trade-off between training costs and mode…
Which Reasoning Trajectories Teach Students to Reason Better? A Simple Metric of Informative Alignment
Yuming Yang, Mingyoung Lai, Wanxu Zhao +13
Long chain-of-thought (CoT) trajectories provide rich supervision signals for distilling reasoning from teacher to student LLMs. However, both prior work and our experiments show t…
Evolution of Concepts in Language Model Pre-Training
Xuyang Ge, Wentao Shu, Jiaxing Wu +3
Language models obtain extensive capabilities through pre-training. However, the pre-training process remains a black box. In this work, we track linear interpretable feature evolu…
Implicit Reward as the Bridge: A Unified View of SFT and DPO Connections
Bo Wang, Qinyuan Cheng, Runyu Peng +7
Post-training processes are essential phases in grounding pre-trained language models to real-world tasks, with learning from demonstrations or preference signals playing a crucial…
Pre-Trained Policy Discriminators are General Reward Models
Shihan Dou, Shichun Liu, Yuming Yang +19
We offer a novel perspective on reward modeling by formulating it as a policy discriminator, which quantifies the difference between two policies to generate a reward signal, guidi…
Capability Salience Vector: Fine-grained Alignment of Loss and Capabilities for Downstream Task Scaling Law
Qiming Ge, Shuhao Xing, Songyang Gao +8
Scaling law builds the relationship between training computation and validation loss, enabling researchers to effectively predict the loss trending of models across different level…
Revisiting the Test-Time Scaling of o1-like Models: Do they Truly Possess Test-Time Scaling Capabilities?
Zhiyuan Zeng, Qinyuan Cheng, Zhangyue Yin +2
The advent of test-time scaling in large language models (LLMs), exemplified by OpenAI's o1 series, has advanced reasoning capabilities by scaling computational resource allocation…
Scaling of Search and Learning: A Roadmap to Reproduce o1 from Reinforcement Learning Perspective
Zhiyuan Zeng, Qinyuan Cheng, Zhangyue Yin +6
OpenAI o1 represents a significant milestone in Artificial Inteiligence, which achieves expert-level performances on many challanging tasks that require strong reasoning ability.Op…
GAOKAO-Eval: Does high scores truly reflect strong capabilities in LLMs?
Zhikai Lei, Tianyi Liang, Hanglei Hu +8
Large Language Models (LLMs) are commonly evaluated using human-crafted benchmarks, under the premise that higher scores implicitly reflect stronger human-like performance. However…
Llama Scope: Extracting Millions of Features from Llama-3.1-8B with Sparse Autoencoders
Zhengfu He, Wentao Shu, Xuyang Ge +9
Sparse Autoencoders (SAEs) have emerged as a powerful unsupervised method for extracting sparse representations from language models, yet scalable training remains a significant ch…
Towards Universality: Studying Mechanistic Similarity Across Language Model Architectures
Junxuan Wang, Xuyang Ge, Wentao Shu +4
The hypothesis of Universality in interpretability suggests that different neural networks may converge to implement similar algorithms on similar tasks. In this work, we investiga…
BitStack: Any-Size Compression of Large Language Models in Variable Memory Environments
Xinghao Wang, Pengyu Wang, Bo Wang +3
Large language models (LLMs) have revolutionized numerous applications, yet their deployment remains challenged by memory constraints on local devices. While scaling laws have enha…
Data Mixing Laws: Optimizing Data Mixtures by Predicting Language Modeling Performance
Jiasheng Ye, Peiju Liu, Tianxiang Sun +3
Pretraining data of large language models composes multiple domains (e.g., web texts, academic papers, codes), whose mixture proportions crucially impact the competence of outcome…
Turn Waste into Worth: Rectifying Top- Router of MoE
Zhiyuan Zeng, Qipeng Guo, Zhaoye Fei +7
Sparse Mixture of Experts (MoE) models are popular for training large language models due to their computational efficiency. However, the commonly used top- routing mechanism su…
Code Needs Comments: Enhancing Code LLMs with Comment Augmentation
Demin Song, Honglin Guo, Yunhua Zhou +8
The programming skill is one crucial ability for Large Language Models (LLMs), necessitating a deep understanding of programming languages (PLs) and their correlation with natural…
Data-free Weight Compress and Denoise for Large Language Models
Runyu Peng, Yunhua Zhou, Qipeng Guo +4
Large Language Models (LLMs) are reshaping the research landscape in artificial intelligence, particularly as model parameters scale up significantly, unlocking remarkable capabili…
AnyGPT: Unified Multimodal LLM with Discrete Sequence Modeling
Jun Zhan, Junqi Dai, Jiasheng Ye +13
We introduce AnyGPT, an any-to-any multimodal language model that utilizes discrete representations for the unified processing of various modalities, including speech, text, images…