The paper introduces the Task Alignment Benchmark (TAB) to evaluate terminal agents' ability to selectively follow relevant environmental instructions while ignoring misleading distractors, revealing…

The paper introduces RedundancyBench, a new benchmark for detecting unnecessary steps in LLM agent trajectories, finding that this task is highly complex and difficult to solve.

This paper investigates various methods for encoding factored tasks, a compact planning representation, into propositional logic for use with SAT solvers, analyzing the impact of encoding choices and…

MAESTRO is a structured pruning framework designed for MoE language models that models autoregressive expert activation trajectories as Ergodic Markov chains, yielding a globally aware importance heur…

A co-designed hardware-software approach for task-conditional sparsity in multi-task inference models, reducing FLOPs, latency, and energy.

TRACE proposes a novel method to mitigate catastrophic forgetting in continual LLM fine-tuning by identifying and isolating a small, task-specific subset of essential parameters for each task.

This paper introduces structured cut sets, a novel preprocessing technique for Maximum k-Cut, and extends existing techniques from Maximum Cut. The rules are optimality-preserving and yield significan…

A framework called DexCompose is proposed to reuse pretrained dexterous policies for multi-task manipulation with explicit finger-level action ownership.

The paper introduces TASTE, an automatic task synthesis method that generates challenging agent benchmarks by evolving tool sequences, demonstrating that existing benchmarks are saturated and that TAS…

ZipRL introduces an adaptive context compression framework that significantly improves the performance and efficiency of LLMs in complex, multi-turn agent tasks by combining multi-granularity compress…

The paper analyzes token reduction for efficient unified VLM training, finding that while task-specific acceleration saves computation, it destroys the mutual performance gains achieved through joint…

The paper proposes SLAT, a segment-level adaptive trimming framework, which efficiently reduces redundant reasoning in large language model CoT outputs by selectively suppressing segments with low mar…

SelSkill introduces a dual-granularity preference learning framework that treats skill use as a 'skill-or-skip' decision, significantly improving agent performance and execution precision in complex a…

The paper proposes FeasiGen, a method to automatically create infeasible tasks for tool-using agents, and finds that most current agents struggle significantly to detect and stop when faced with such…

CORE-MTL proposes a representation-centric framework that uses causal orthogonal representations to disentangle task-relevant structure from nuisance variation in multi-task learning, achieving superi…

The paper proposes ReuseRL, a method that improves agent generalization in Reinforcement Learning by enforcing structural compressibility of successful agent trajectories into reusable skills.

This paper introduces MCTS-Guided Group Relative Policy Optimization (M-GRPO) to enhance LLM spatial reasoning by improving the decomposition of complex tasks into optimal sub-tasks.

The paper demonstrates that extended pure neural reasoning fails on complex, deterministic state-tracking tasks beyond a certain 'Deterministic Horizon,' necessitating the integration of external tool…

DeepTool introduces a novel Process-Supervised Reinforcement Learning framework to enhance Tool-Integrated Reasoning by explicitly supervising and rewarding intermediate, interleaved deliberation step…