91 citations · 270 across the 62 of their papers we have counts for
66 papers
Learning to Extrapolate in Depth with Looped Transformers
Huitao Yang, Jianhao Huang, Zixuan Wang +1
Looped transformers reuse a single block across iterations, raising a natural question: how far can a block trained at one depth be safely extrapolated to greater depths? We study…
Distilling Visual Reasoning into Text Space
Wenhan Yang, Nilay Naharas, Ali Payani +1
Large Vision-Language Models (LVLMs) have shown strong promise for multimodal reasoning, yet often struggle with tasks requiring concepts beyond what is directly observable in the…
Locating Hidden Failures Makes Long-Horizon Agents More Reliable
Salman Rahman, Yubin Kim, Mihir Parmar +15
As AI agents take on long, autonomous tasks, we increasingly oversee rather than perform the work, yet we still judge them almost entirely by whether they finally succeed. An outco…
Learning as Reasoning Unfolds: Progressive Rollout Allocation for Efficient Reinforcement Learning
Heyang Jiang, Henry Liu, Baharan Mirzasoleiman
Reinforcement learning with verifiable rewards (RLVR) has emerged as a highly effective framework for improving LLM reasoning, with methods such as GRPO among its most successful i…
Reasoning Quality Emerges Early: Data Curation for Reasoning Models
Hongyi Henry Jin, Wenhan Yang, Meysam Ghaffari +2
Supervised fine-tuning (SFT) on a small, high-quality set of long reasoning traces is an effective approach for eliciting strong reasoning capabilities in Large Language Models (LL…
Your Model Already Knows: Attention-Guided Safety Filter for Vision-Language-Action Models
Seongbin Park, Fan Zhang, Baharan Mirzasoleiman +2
Vision-Language-Action (VLA) models have demonstrated impressive end-to-end performance across a variety of robotic manipulation tasks. However, these policies offer no guarantees…
ProbeAct: Probe-Guided Training-Free Failure Recovery in Vision-Language-Action Models
Fan Zhang, Seongbin Park, Baharan Mirzasoleiman +2
Vision-Language-Action (VLA) models demonstrate strong perfor-1 mance on language-conditioned robotic manipulation within their training dis-2 tribution, yet their generalization c…
How Transformers Learn to Plan via Multi-Token Prediction
Jianhao Huang, Zhanpeng Zhou, Renqiu Xia +3
While next-token prediction (NTP) has been the standard objective for training language models, it often struggles to capture global structure in reasoning tasks. Multi-token predi…
Theoretical Perspectives on Data Quality and Synergistic Effects in Pre- and Post-Training Reasoning Models
Adel Javanmard, Baharan Mirzasoleiman, Vahab Mirrokni
Large Language Models (LLMs) are pretrained on massive datasets and later instruction-tuned via supervised fine-tuning (SFT) or reinforcement learning (RL). Best practices emphasiz…
Beyond What Seems Necessary: Hidden Gains from Scaling Training-Time Reasoning Length under Outcome Supervision
Yihao Xue, Allan Zhang, Jianhao Huang +2
Training LLMs to think and reason for longer has become a key ingredient in building state-of-the-art models that can solve complex problems previously out of reach. Recent efforts…
Data Distribution as a Lever for Guiding Optimizers Toward Superior Generalization in LLMs
Tushaar Gangavarapu, Jiping Li, Christopher Vattheuer +2
Can modifying the training data distribution guide optimizers toward solutions with improved generalization when training large language models (LLMs)? In this work, we theoretical…
Tuning the Implicit Regularizer of Masked Diffusion Language Models: Enhancing Generalization via Insights from -Parity
Jianhao Huang, Baharan Mirzasoleiman
Masked Diffusion Language Models have recently emerged as a powerful generative paradigm, yet their generalization properties remain understudied compared to their auto-regressive…
Understanding the Role of Training Data in Test-Time Scaling
Adel Javanmard, Baharan Mirzasoleiman, Vahab Mirrokni
Test-time scaling improves the reasoning capabilities of large language models (LLMs) by allocating extra compute to generate longer Chains-of-Thoughts (CoTs). This enables models…
Data Selection for Fine-tuning Vision Language Models via Cross Modal Alignment Trajectories
Nilay Naharas, Dang Nguyen, Nesihan Bulut +3
Data-efficient learning aims to eliminate redundancy in large training datasets by training models on smaller subsets of the most informative examples. While data selection has bee…
LoRA is All You Need for Safety Alignment of Reasoning LLMs
Yihao Xue, Baharan Mirzasoleiman
Reasoning-capable LLMs have achieved major breakthroughs in solving complex problems, but recent work shows that acquiring and deploying strong reasoning can introduce significant…
Beyond Semantic Entropy: Boosting LLM Uncertainty Quantification with Pairwise Semantic Similarity
Dang Nguyen, Ali Payani, Baharan Mirzasoleiman
Hallucination in large language models (LLMs) can be detected by assessing the uncertainty of model outputs, typically measured using entropy. Semantic entropy (SE) enhances tradit…
Bootstrapping LLM Robustness for VLM Safety via Reducing the Pretraining Modality Gap
Wenhan Yang, Spencer Stice, Ali Payani +1
Ensuring Vision-Language Models (VLMs) generate safe outputs is crucial for their reliable deployment. However, LVLMs suffer from drastic safety degradation compared to their LLM b…
Do We Need All the Synthetic Data? Targeted Image Augmentation via Diffusion Models
Dang Nguyen, Jiping Li, Jinghao Zheng +1
Synthetically augmenting training datasets with diffusion models has become an effective strategy for improving the generalization of image classifiers. However, existing approache…
DD-Ranking: Rethinking the Evaluation of Dataset Distillation
Zekai Li, Xinhao Zhong, Samir Khaki +49
In recent years, dataset distillation has provided a reliable solution for data compression, where models trained on the resulting smaller synthetic datasets achieve performance co…
Synthetic Text Generation for Training Large Language Models via Gradient Matching
Dang Nguyen, Zeman Li, Mohammadhossein Bateni +3
Synthetic data has the potential to improve the performance, training efficiency, and privacy of real training examples. Nevertheless, existing approaches for synthetic text genera…
Verify when Uncertain: Beyond Self-Consistency in Black Box Hallucination Detection
Yihao Xue, Kristjan Greenewald, Youssef Mroueh +1
Large Language Models (LLMs) often hallucinate, limiting their reliability in sensitive applications. In black-box settings, several self-consistency-based techniques have been pro…
Representations Shape Weak-to-Strong Generalization: Theoretical Insights and Empirical Predictions
Yihao Xue, Jiping Li, Baharan Mirzasoleiman
Weak-to-Strong Generalization (W2SG), where a weak model supervises a stronger one, serves as an important analogy for understanding how humans might guide superhuman intelligence…
MM-GEN: Enhancing Task Performance Through Targeted Multimodal Data Curation
Siddharth Joshi, Besmira Nushi, Vidhisha Balachandran +4
Vision-language models (VLMs) are highly effective but often underperform on specialized tasks; for example, Llava-1.5 struggles with chart and diagram understanding due to scarce…
Dataset Distillation via Knowledge Distillation: Towards Efficient Self-Supervised Pre-Training of Deep Networks
Siddharth Joshi, Jiayi Ni, Baharan Mirzasoleiman
Dataset distillation (DD) generates small synthetic datasets that can efficiently train deep networks with a limited amount of memory and compute. Despite the success of DD methods…