… View arXiv page View PDF Project page Add to collection Community Reinforcement learning RL has substantially improved the ability of large language model LLM agents to interact with environments and solve multi-turn tasks. …
… View arXiv page View PDF Project page Add to collection Community Unified Multimodal Models UMMs have emerged as a critical direction for general-purpose multimodal intelligence, integrating understanding and generation into a single framework. …
… View arXiv page View PDF GitHub 149 Add to collection Community This is an automated message from the Librarian Bot . …
… View arXiv page View PDF GitHub 1 Add to collection Community token credit assignment for reinforcement learning based on guidance signals This is an automated message from the Librarian Bot . …
… View arXiv page View PDF GitHub 2 Add to collection Community Standard OPD may under-cover plausible alternatives, while EOPD relies solely on teacher entropy and uncalibrated targets. …
… View arXiv page View PDF Project page GitHub 7 Add to collection Community AVTok: 1D Unified Tokenization for Holistic Audio-Video Generation This is an automated message from the Librarian Bot . …
… View arXiv page View PDF Add to collection Community Hi everyone! …
… View arXiv page View PDF Project page GitHub 7 Add to collection Community Codes & Datasets: https://github.com/Graph-COM/TurnGate Project Website: https://turn-gate.github.io/ Arxiv: https://arxiv.org/abs/2605.05630 TurnGate is a response-aware defense mechanism designed to detect and mitigate hid… …
… Project Page: https://shjo-april.github.io/ Chain-of-Prompts / View arXiv page View PDF Project page GitHub 8 Add to collection Community 🔥 Early Accepted to MICCAI 2026 Top 9% TL;DR: We introduce Chain-of-Prompts CoP , a training-free framework that shifts the interaction cost for cell instance se… …
… View arXiv page View PDF Project page GitHub 34 Add to collection Community MARBLE harmonizes reward-specific policy gradients into a single update direction, simultaneously improving all rewards in one training run. …