ARBOR introduces a reusable rubric buffer that consolidates contrastive trajectory drafts into cross-query rubrics for online process rewards, outperforming GRPO and DAPO on multi-hop QA benchmarks.
Agent workflow memory
5 Pith papers cite this work, alongside 1 external citations. Polarity classification is still indexing.
5
Pith papers citing it
1
external citations · OpenAlex
years
2026 5representative citing papers
CurateEvo evolves executable data-curation code using failed agent trajectories, improving post-training performance by 3.2 and 2.7 points over baselines on labeled and wild data respectively.
MosaicKV achieves up to 16x attention speedup, 4.8x lower decode latency, 7.3x higher throughput, and 3x memory reduction with 1.76% accuracy loss via dynamic two-D KV cache compression and management on H800 GPUs.
Mem-π is a framework using a dedicated model and decision-content decoupled RL to generate context-specific guidance on demand for LLM agents, outperforming retrieval baselines by over 30% on web navigation.
ReVision reduces token usage by 46% and improves success rate by 3% on OSWorld, WebTailBench, and AgentNetBench by removing redundant visual patches from 5-history trajectories with Qwen2.5-VL-7B.
citing papers explorer
-
ARBOR: Online Process Rewards via a Reusable Rubric Buffer for Search Agents
ARBOR introduces a reusable rubric buffer that consolidates contrastive trajectory drafts into cross-query rubrics for online process rewards, outperforming GRPO and DAPO on multi-hop QA benchmarks.
-
CurateEvo: Data-Curation Evolving for Agentic Post-Training
CurateEvo evolves executable data-curation code using failed agent trajectories, improving post-training performance by 3.2 and 2.7 points over baselines on labeled and wild data respectively.
-
MosaicKV: Serving Long-Context LLM with Dynamic Two-D KV Cache Compression
MosaicKV achieves up to 16x attention speedup, 4.8x lower decode latency, 7.3x higher throughput, and 3x memory reduction with 1.76% accuracy loss via dynamic two-D KV cache compression and management on H800 GPUs.
-
Mem-$\pi$: Adaptive Memory through Learning When and What to Generate
Mem-π is a framework using a dedicated model and decision-content decoupled RL to generate context-specific guidance on demand for LLM agents, outperforming retrieval baselines by over 30% on web navigation.
-
ReVision: Scaling Computer-Use Agents via Temporal Visual Redundancy Reduction
ReVision reduces token usage by 46% and improves success rate by 3% on OSWorld, WebTailBench, and AgentNetBench by removing redundant visual patches from 5-history trajectories with Qwen2.5-VL-7B.