EgoMemReason is a new benchmark showing that even the best multimodal models achieve only 39.6% accuracy on reasoning tasks that require integrating sparse evidence across days in egocentric video.
hub
Transactions of the association for computational linguistics , volume=
10 Pith papers cite this work. Polarity classification is still indexing.
10
Pith papers citing it
hub tools
years
2026 10representative citing papers
GraphBit is a DAG-based engine-orchestrated framework for agentic LLMs that achieves 67.6% accuracy with zero hallucinations on GAIA benchmarks.
OnePred maintains a recursively updated intent memory and uses two-stage RL to predict next queries, cutting token use by up to 22x while outperforming baselines on a new NQP-Bench dataset.
GoLongRL releases a 23K-sample open long-context RL dataset spanning 9 tasks and introduces TMN-Reweight to improve multitask optimization, achieving performance comparable to much larger models under GRPO.
Conflicting biomedical evidence triggers order-dependent prediction flips in RAG LLMs, and a new abstention score combining confidence with conflict detection raises selective accuracy by 7-33 points in the hardest conditions.
Attention to goal tokens declines in multi-turn LLM interactions while residual representations often retain decodable goal information, and the gap between these predicts whether goal-conditioned behavior survives.
ProCompNav builds a candidate pool from ambiguous queries then uses pool-splitting binary questions for disambiguation, improving success rate and shortening responses on CoIN-Bench and TextNav.
APPS approximates power targets p(x)^alpha via parallel particle propagation with proposal-corrected reweighting and future-value-guided selection at block boundaries, improving accuracy-runtime trade-offs in training-free decoding.
Re-analysis of MCP-grounded CTF agent runs shows Skills add only 8.9 pp (p=0.71) when tool feedback is strict and low-latency, supporting the hypothesis that high environment-feedback bandwidth reduces the marginal value of curated procedural knowledge.
UserGPT introduces a generative LLM framework with a behavior simulation engine, semantization module, and DF-GRPO post-training that scores 0.7325 on tag prediction and 0.7528 on summary generation on HPR-Bench while compressing records by up to 97.9%.
citing papers explorer
-
EgoMemReason: A Memory-Driven Reasoning Benchmark for Long-Horizon Egocentric Video Understanding
EgoMemReason is a new benchmark showing that even the best multimodal models achieve only 39.6% accuracy on reasoning tasks that require integrating sparse evidence across days in egocentric video.
-
GraphBit: A Graph-based Agentic Framework for Non-Linear Agent Orchestration
GraphBit is a DAG-based engine-orchestrated framework for agentic LLMs that achieves 67.6% accuracy with zero hallucinations on GAIA benchmarks.
-
OnePred: Next-Query Prediction via Recursive Intent Memory in Multi-Turn Conversations
OnePred maintains a recursively updated intent memory and uses two-stage RL to predict next queries, cutting token use by up to 22x while outperforming baselines on a new NQP-Bench dataset.
-
GoLongRL: Capability-Oriented Long Context Reinforcement Learning with Multitask Alignment
GoLongRL releases a 23K-sample open long-context RL dataset spanning 9 tasks and introduces TMN-Reweight to improve multitask optimization, achieving performance comparable to much larger models under GRPO.
-
When Evidence Conflicts: Uncertainty and Order Effects in Retrieval-Augmented Biomedical Question Answering
Conflicting biomedical evidence triggers order-dependent prediction flips in RAG LLMs, and a new abstention score combining confidence with conflict detection raises selective accuracy by 7-33 points in the hardest conditions.
-
When Attention Closes: How LLMs Lose the Thread in Multi-Turn Interaction
Attention to goal tokens declines in multi-turn LLM interactions while residual representations often retain decodable goal information, and the gap between these predicts whether goal-conditioned behavior survives.
-
ProCompNav: Proactive Instance Navigation with Comparative Judgment for Ambiguous User Queries
ProCompNav builds a candidate pool from ambiguous queries then uses pool-splitting binary questions for disambiguation, improving success rate and shortening responses on CoIN-Bench and TextNav.
-
The Model Knows, the Decoder Finds: Future Value Guided Particle Power Sampling
APPS approximates power targets p(x)^alpha via parallel particle propagation with proposal-corrected reweighting and future-value-guided selection at block boundaries, improving accuracy-runtime trade-offs in training-free decoding.
-
When Skills Don't Help: A Negative Result on Procedural Knowledge for Tool-Grounded Agents in Offensive Cybersecurity
Re-analysis of MCP-grounded CTF agent runs shows Skills add only 8.9 pp (p=0.71) when tool feedback is strict and low-latency, supporting the hypothesis that high environment-feedback bandwidth reduces the marginal value of curated procedural knowledge.
-
UserGPT Technical Report
UserGPT introduces a generative LLM framework with a behavior simulation engine, semantization module, and DF-GRPO post-training that scores 0.7325 on tag prediction and 0.7528 on summary generation on HPR-Bench while compressing records by up to 97.9%.