3 papers
cs.LG2026
TriRoute: Unified Learned Routing for Joint Adaptive Attention, Experts, and KV-Cache Allocation
Andrii Balashov, Olena Ponomarova
Conditional computation can decouple language model quality from per-token inference cost, yet leading techniques act on a single axis in isolation: Mixture-of-Experts (MoE) sparsi…
cs.LG2025
Reinforcement Learning Fine-Tunes a Sparse Subnetwork in Large Language Models
Andrii Balashov
Reinforcement learning (RL) is a key post-pretraining step for aligning large language models (LLMs) with complex tasks and human preferences. While it is often assumed that RL fin…
cs.CR2025
Multi-Stage Prompt Inference Attacks on Enterprise LLM Systems
Andrii Balashov, Olena Ponomarova, Xiaohua Zhai
Large Language Models (LLMs) deployed in enterprise settings (e.g., as Microsoft 365 Copilot) face novel security challenges. One critical threat is prompt inference attacks: adver…