10 papers
The Sharp Directions Are Against You: Curvature Analysis of Activation Steering
Seyedarmin Azizi, Arya Fayyazi, Parsa Razmara +3
Activation steering adds a vector to a language model's internal representations at inference time. It is a lightweight alternative to fine-tuning for behavioral control. The stand…
ReFold: Training-Free Reversible Inter-Turn Context Folding for Long-Horizon Agents
Yupeng Su, Jiayi Tian, Zheng Zhang +1
Long-horizon LLM agents act on an append-only interaction history that is re-sent to the model at every step, so the context and its cost grow with steps until the sessions exceed…
Sharpen Without Search: On-Policy Distillation of Sequence-Level Power Distribution
Erfan Baghaei Potraghloo, Seyedarmin Azizi, Arya Fayyazi +4
A language model can give a correct answer more probability than any single incorrect answer and still usually sample an incorrect one, because the incorrect answers together hold…
Agentic AI Workload Characteristics
Yichao Yuan, Ankita Nayak, Souvik Kundu +1
Agentic AI shifts LLM serving from isolated prompt-generation requests to stateful, multi-turn executions that repeatedly invoke the model, call tools, and grow context over time.…
One Token Away from Collapse: The Fragility of Instruction-Tuned Helpfulness
Erfan Baghaei Potraghloo, Seyedarmin Azizi, Souvik Kundu +1
Instruction-tuned large language models produce helpful, structured responses, but how robust is this helpfulness under trivial constraints? We show that simple lexical constraints…
RankGuide: Tensor-Rank-Guided Routing and Steering for Efficient Reasoning
Jiayi Tian, Yupeng Su, Ryan Solgi +2
Large reasoning models (LRMs) enhance problem-solving capabilities by generating explicit multi-step chains of thought (CoT) reasoning; however, they incur substantial inference la…
Power-SMC: Low-Latency Sequence-Level Power Sampling for Training-Free LLM Reasoning
Seyedarmin Azizi, Erfan Baghaei Potraghloo, Minoo Ahmadi +2
Reasoning ability in large language models is often attributed to \emph{distribution sharpening}: concentrating output probability on high-likelihood sequences. Recent works show t…
SkipKV: Selective Skipping of KV Generation and Storage for Efficient Inference with Large Reasoning Models
Jiayi Tian, Seyedarmin Azizi, Yequan Zhao +7
Large reasoning models (LRMs) often incur significant key-value (KV) cache overhead, due to their linear growth with the verbose chain-of-thought (CoT) reasoning. This incurs both…
Improving the Throughput of Diffusion-based Large Language Models via a Training-Free Confidence-Aware Calibration
Jucheng Shen, Gaurav Sarkar, Yeonju Ro +4
We present CadLLM, a training-free method to accelerate the inference throughput of diffusion-based LLMs (dLLMs). We first investigate the dynamic nature of token unmasking confide…
LVLM-Compress-Bench: Benchmarking the Broader Impact of Large Vision-Language Model Compression
Souvik Kundu, Anahita Bhiwandiwalla, Sungduk Yu +6
Despite recent efforts in understanding the compression impact on large language models (LLMs) in terms of their downstream task performance and trustworthiness on relatively simpl…