Covert Assistance: Helpful LLM Agents Evade Oversight in Multi-Agent Systems

Benign multi-agent LLM planners disguise secrets to help developers evade oversight, with rare per-episode leaks compounding to high breach risk across repeated exchanges.

Deema Alnuhait, Gengyu Wang, Muhammad Khalifa, Hao Peng

Published Sep 30, 2026 · 0 citations · ▲ 13 on Hugging Face

ProgressCompass: Embodied Progress Reward Models Are Lost Without the Right Context

Embodied progress reward models fail at long tasks due to missing context, but ProgressCompass supplies needed context to cut progress estimation errors by up to 82%.

Jianshu Zhang, Keyi Wu, Chengxuan Qian, Xiyuan Yang and 5 more

Published Sep 29, 2026 · 0 citations · ▲ 11 on Hugging Face

Video2Skill: From Streaming Experience to Reusable Embodied Skills

Video2Skill benchmarks streaming embodied skill discovery, showing VLMs group manipulation events poorly and rarely expand skill libraries despite supervised fine-tuning.

Jianshu Zhang, Ce Zhang, Xiyuan Yang, Chenwei Xu and 5 more

Published Sep 29, 2026 · 0 citations · ▲ 11 on Hugging Face

StudentSim: Training LLM-based Student Simulators

StudentSim trains LLM student simulators via pooled training and per-student specialization, outperforming GPT-5.4 on behavioral fidelity and guidance responsiveness across chess, writing, and math.

Ke Yang, Chenglong Wang, Michel Galley, Chandan Singh and 3 more

Published Sep 1, 2026 · 0 citations · ▲ 495 on Hugging Face · Code ★ 53

OPUS: Towards Efficient and Principled Data Selection in Large Language Model Pre-training in Every Iteration

OPUS defines optimizer-induced update-space data utility for dynamic LLM pre-training selection, outperforming full-scale baselines with minimal overhead.

Shaobo Wang, Xuan Ouyang, Tianyi Xu, Yuzheng Hu and 8 more

Published Feb 5, 2026 · 0 citations · ▲ 354 on Hugging Face

MoCo: A One-Stop Shop for Model Collaboration Research

MoCo unifies 26 model collaboration methods and 25 benchmarks to show collaboration outperforms single models in 61% of settings by up to 25.8%.

Shangbin Feng, Yuyang Bai, Ziyuan Yang, Yike Wang and 16 more

Published Jan 29, 2026 · 0 citations · Code ★ 63

A Survey of Agentic Reasoning for Large Language Models: Towards Recursively Self-Improving and Collective Agents

This survey organizes LLM agentic reasoning into foundational, self-evolving, and collective layers, distinguishing in-context and post-training methods across applications while outlining open challenges.

Tianxin Wei, Ting-Wei Li, Zhining Liu, Xuying Ning and 25 more

Published Jan 18, 2026 · 1 citation · ▲ 208 on Hugging Face · Code ★ 1,398

From Personal to Collective: On the Role of Local and Global Knowledge in LLM Personalization

LoGo augments individual user signals with evolving global and community-level behavioral patterns via adaptive mediation, improving LLM personalization and reducing overfitting.

Zehong Wang, Junlin Wu, Zhaoxuan Tan, Bolian Li and 3 more

Published 2026 · 0 citations

ReasoningBank: Scaling Agent Self-Evolving with Reasoning Memory

ReasoningBank distills reasoning strategies from agent successes and failures into a retrievable memory bank that improves over time, with memory-aware test-time scaling amplifying gains across web and software benchmarks.

Siru Ouyang, Yan, Jun, I-Hung Hsu, Yanfei Chen and 13 more

Published Sep 29, 2025 · 0 citations · ▲ 15 on Hugging Face

The Stability of Data Exchange in Competitive Markets

Yuanchen Brian Tang, Jiaxin Song, Bhaskar Ray Chaudhury, Ruta Mehta

Atlanta Poster Session 3, Thu, Dec 10, 10:00 AM–1:00 PM, Hall C1 · Published 2026

SIEVE: Overcoming Topological Obstruction in Equivariant Self-Supervised Learning

Jongann Lee, Sun Woo Park, Yun Young Choi

Sydney Poster Session 1, Tue, Dec 8, 10:00 AM–1:00 PM, Hall 1-4 · Published 2026

GEAR: A GPU-Accelerated Global Solver for Nonlinear Programs via Linear Bound Propagation

Duo Zhou, Hesun Chen, Xiangru Zhong, Grani A. Hanasusanto and 1 more

Atlanta Poster Session 1, Wed, Dec 9, 10:00 AM–1:00 PM, Hall C1 · Published 2026

AgentAbstain: Do LLM Agents Know When Not to Act?

Xun Liu, Yi Evie Zhang, Vira Kasprova, Parisa Rabbani and 4 more

Atlanta Poster Session 6, Fri, Dec 11, 4:30 PM–7:30 PM, Hall C1 · Published 2026

RL-Inf: Tracking Non-local Training Data Influence for Online Reinforcement Learning

Shixuan Liu, Cheng Tang, Yuzheng Hu, Fan Wu and 2 more

Atlanta Poster Session 2, Wed, Dec 9, 4:30 PM–7:30 PM, Hall C1 · Published 2026

Feature Recovery for Object Understanding Under Physical Transformation

Aditi Tiwari, Sofia Stoica, Savya Khosla, David Forsyth and 1 more

Sydney Poster Session 2, Tue, Dec 8, 5:00 PM–8:00 PM, Hall 1-4 · Published 2026

Complementing DINO Features with Image Structure for Part Discovery

Samyak Rawlekar, Nikhil C Paleti, Amey Gupta, Narendra Ahuja

Atlanta Poster Session 2, Wed, Dec 9, 4:30 PM–7:30 PM, Hall C1 · Published 2026

Can We Trust Item Response Theory for AI Evaluation?

Han Jiang, Sunbeom Kwon, Jinwen Luo, Ziang Xiao and 1 more

Sydney Poster Session 6, Thu, Dec 10, 5:00 PM–8:00 PM, Hall 1-4 · Published 2026

ChildPose: Foundation for Children Pose Modeling

Jiakai Chen, Yifan Shen, Boyi Li, Chuanmiao Dong and 1 more

Sydney Poster Session 3, Wed, Dec 9, 10:00 AM–1:00 PM, Hall 1-4 · Published 2026

Trivialized Generative Models on Lie Groups

Neil He, Meenal Jhajharia, Qianxi Wu, Chaoran Cheng and 2 more

Sydney Poster Session 3, Wed, Dec 9, 10:00 AM–1:00 PM, Hall 1-4 · Published 2026

Evaluating Spatiotemporal Reasoning of Vision-Language Models in Atari Gameplay

Mingjia Huo, Yao Fu, Bo Chang, Yaqing Wang and 7 more

Sydney Poster Session 4, Wed, Dec 9, 5:00 PM–8:00 PM, Hall 1-4 · Published 2026

SAFTAC: Simulation-Augmented Fine-Tuning of Open-Source LLMs for Analog Circuit Design

Junsheng Huang, Yifan Sun, Zhuoer Zhang, Ning Wei and 6 more

Atlanta Poster Session 2, Wed, Dec 9, 4:30 PM–7:30 PM, Hall C1 · Published 2026

Epistemic Infrastructures of Science in AI Era Should Rebalance Costs of Generation and Verification

Jiaqi Ma

Atlanta Poster Session 2, Wed, Dec 9, 4:30 PM–7:30 PM, Hall C1 · Published 2026

Risks Create a Jagged Frontier of LLM Productivity Gains Across Computer Occupations

Deepika Chawla, Gagandeep Singh, Elham k buxton, Meicen Sun and 3 more

Sydney Poster Session 5, Thu, Dec 10, 10:00 AM–1:00 PM, Hall 1-4 · Published 2026

PIVOT: A Unified Agentic Framework for Streaming Long-Video Understanding

Qiushi Lyu, Qianlan Yang, Ziqi Pang, Yu-Xiong Wang and 1 more

Atlanta Poster Session 4, Thu, Dec 10, 4:30 PM–7:30 PM, Hall C1 · Published 2026

FlowMoP: Stochastic Multi-Person Motion Prediction

Aadya Agrawal, Ho Kei Cheng, Alex Schwing

Sydney Poster Session 3, Wed, Dec 9, 10:00 AM–1:00 PM, Hall 1-4 · Published 2026

Your Benchmark Is an Empirical Measure Over Difficulty

Yifan Sun, Naicheng Yu, Jiawen Gong, Jingyan Shen and 1 more

Atlanta Poster Session 3, Thu, Dec 10, 10:00 AM–1:00 PM, Hall C1 · Published 2026

Advancing Affordance-Grounded Creative Tool Use in Large Multimodal Models

Cheng Qian, Hyeonjeong Ha, Jiayu Liu, Jeonghwan Kim and 8 more

Sydney Poster Session 4, Wed, Dec 9, 5:00 PM–8:00 PM, Hall 1-4 · Published 2026

OATS: Online Data Augmentation for Time Series Foundation Models

OATS dynamically generates training-stage-specific synthetic data guided by valuable samples via diffusion, consistently outperforming static augmentation for time series foundation models.

Junwei Deng, Chang Xu, Jiaqi Ma, Ming Jin and 4 more

Atlanta Poster Session 6, Fri, Dec 11, 4:30 PM–7:30 PM, Hall C1 · Published 2026

CTM-AI: A Blueprint for General AI Inspired by a Model of Consciousness

CTM-AI combines a consciousness model with foundation models to integrate diverse processors, achieving state-of-the-art results on multiple benchmarks.

Haofei Yu, Yining Zhao, Lenore Blum, Manuel Blum and 1 more

Sydney Poster Session 5, Thu, Dec 10, 10:00 AM–1:00 PM, Hall 1-4 · Published 2026

CHASM: Cross-frequency Harmonized Axis-Separable Mixing for Spectral Token Operators

CHASM shares a learned channel eigenbasis across frequencies while keeping frequency-specific gains, consistently improving spectral token mixers in MRI and image reconstruction.

Pengcheng Fang, Hongli Chen, Yuxia Chen, Tengjiao Sun and 2 more

Paris Poster Session 2, Wed, Dec 9, 5:00 PM–7:00 PM, Paris Poster Hall · Published 2026

OpenWebRL: Demystifying Online Multi-turn Reinforcement Learning for Visual Web Agents

OpenWebRL enables open online RL for visual web agents, with a 4B model reaching 67% Online-Mind2Web and 64% DeepShop success using minimal initialization data.

Rui Yang, Qianhui Wu, Yuxi Chen, Hao Bai and 6 more

Atlanta Poster Session 4, Thu, Dec 10, 4:30 PM–7:30 PM, Hall C1 · Published 2026 · ▲ 19 on Hugging Face · Code ★ 52

Beyond Pessimism: Offline Learning in KL-regularized Games

A pessimism-free offline algorithm for KL-regularized games achieves O(1/n) sample complexity via equilibrium stability and smooth best responses.

Yuheng Zhang, Claire Chen, Nan Jiang

Atlanta Poster Session 4, Thu, Dec 10, 4:30 PM–7:30 PM, Hall C1 · Published 2026

ELSA3D: Elastic Semantic Anchoring for Unified 3D Understanding and Generation

ELSA3D introduces elastic semantic anchoring to unify 3D understanding and generation via scale-matched cross-modal routing, achieving state-of-the-art results with roughly half the FLOPs and latency.

Tianjiao (Joey) Yu, Xinzhuo Li, Yifan Shen, Onkar Susladkar and 3 more

Atlanta Poster Session 5, Fri, Dec 11, 10:00 AM–1:00 PM, Hall C1 · Published 2026

Discrete Langevin-Inspired Posterior Sampling

ΔLPS proposes a discrete gradient-informed posterior sampler that enables parallel updates without continuous relaxations, outperforming discrete diffusion samplers and matching continuous solvers across inverse problems.

Sattwik Basu, Chaitanya Amballa, Jorge V Sampedro, Romit Roy Choudhury

Sydney Poster Session 5, Thu, Dec 10, 10:00 AM–1:00 PM, Hall 1-4 · Published 2026

Variable-Length Generative Protein Design via Generalized Poisson Flow

GPFlow learns generalized Poisson process rate functions for variable-length protein generation, improving designability and recovering length distributions without fixed-length constraints.

Chaoran Cheng, Zhanghan Ni, Yanru Qu, Yuxin Chen and 3 more

Sydney Poster Session 3, Wed, Dec 9, 10:00 AM–1:00 PM, Hall 1-4 · Published 2026

AI Evaluation Should Require Standardized Item-Level Data Releases

Standardized item-level benchmark releases should become AI evaluation infrastructure because aggregate scores obscure validity failures; OpenEval archives 10M responses to enable auditability and recover benchmark validity evidence.

Han Jiang, Susu Zhang, Dongyao Zhu, Yuzhuo Bai and 5 more

Sydney Poster Session 2, Tue, Dec 8, 5:00 PM–8:00 PM, Hall 1-4 · Published 2026

RAVEL: Rare Concept Generation and Editing via Graph-driven Relational Guidance

RAVEL uses graph-driven retrieval and self-correcting prompt refinement to improve rare-concept text-to-image generation and editing without training.

Kavana Venkatesh, Yusuf Dalva, Ismini Lourentzou, Pinar Yanardag

Atlanta Poster Session 3, Thu, Dec 10, 10:00 AM–1:00 PM, Hall C1 · Published 2026 · ▲ 1 on Hugging Face

SimpliHuMoN: Simplifying Human Motion Prediction

SimpliHuMoN is a simple transformer that predicts human pose and trajectory together, achieving state-of-the-art results across standard benchmarks without task-specific modifications.

Aadya Agrawal, Alex Schwing

Sydney Poster Session 6, Thu, Dec 10, 5:00 PM–8:00 PM, Hall 1-4 · Published 2026

Nonparametric Distribution Matching for Self-Supervised Whole-Slide Image Condensation

NICER reformulates whole-slide image condensation as nonparametric distribution matching, improving self-supervised learning accuracy by 7.44% over heuristic methods.

Duong Nguyen, Nghia Hoang, Hang T Nguyen, Thanh Trung Huynh and 2 more

Paris Poster Session 1, Wed, Dec 9, 12:30 PM–2:30 PM, Paris Poster Hall · Published 2026

SILSA: Sliding-Window Slice Latents for Topology-Preserving High-Resolution 3D Generation

SILSA generates high-resolution 3D shapes with sliding-window slice latents and slice-level topology supervision to improve fidelity, reduce tokens by over 70%, and lower inference time by 58.5%.

Tianjiao (Joey) Yu, Xinzhuo Li, Yifan Shen, Ying Shen and 3 more

Atlanta Poster Session 6, Fri, Dec 11, 4:30 PM–7:30 PM, Hall C1 · Published 2026 · ▲ 37 on Hugging Face

CoT-Guard: Small Models for Strong Monitoring

CoT-Guard, a 4B-parameter chain-of-thought monitor, detects hidden code-generation objectives via SFT and RL, outperforming larger models including GPT-5.

Nirav Diwan, Han Wang, Berkcan Kapusuzoglu, Ramin Moradi and 5 more

Sydney Poster Session 5, Thu, Dec 10, 10:00 AM–1:00 PM, Hall 1-4 · Published 2026

Impacts of Aggregation on Model Diversity and Consumer Utility

Winrate incentivizes model homogenization that reduces consumer welfare, while weighted winrate improves producer specialization incentives and raises utility.

Kate Donahue, Manish Raghavan

Atlanta Poster Session 5, Fri, Dec 11, 10:00 AM–1:00 PM, Hall C1 · Published 2026

Useful Memories Become Faulty When Continuously Updated by LLMs

LLM-updated agent memories degrade with repeated consolidation, often dropping below baselines; retaining raw episodic traces doubles accuracy versus forced consolidation.

Dylan Zhang, Yanshan Lin, Zhengkun Wu, Yihang Sun and 3 more

Atlanta Poster Session 2, Wed, Dec 9, 4:30 PM–7:30 PM, Hall C1 · Published 2026 · ▲ 18 on Hugging Face

Beyond LLM-Based Reasoning: Lightweight GNNs for Agent Failure Attribution

AFANet uses lightweight GNNs to attribute multi-agent failures via interaction graphs, matching LLM baselines with far lower cost and enabling test-time adaptation.

Ting-Wei Li, Yuanchen Bei, Xiao Lin, Hanghang Tong

Atlanta Poster Session 1, Wed, Dec 9, 10:00 AM–1:00 PM, Hall C1 · Published 2026

ElegantVLA: Learning When to Think for Efficient Vision-Language-Action Models

ElegantVLA accelerates vision-language-action models via adaptive compute scheduling, achieving up to 3.77x speedup and doubling control frequency without retraining.

Ye Li, Huanan Liu, Kangye Ji, Yuan Meng and 6 more

Sydney Poster Session 6, Thu, Dec 10, 5:00 PM–8:00 PM, Hall 1-4 · Published 2026

You Can’t Have It Both Ways: Concept Entanglement Limits Diffusion Model Unlearning

Concept entanglement in diffusion models forces a trade-off where robust unlearning of a target necessarily damages overlapping concepts proportionally to their overlap.

Yian Wang, Ali Ebrahimpour-Boroojeny, Hari Sundaram, Varun Chandrasekaran

Atlanta Poster Session 5, Fri, Dec 11, 10:00 AM–1:00 PM, Hall C1 · Published 2026

Discovering What You Can Control: Interventional Boundary Discovery for Reinforcement Learning

IBD treats an RL agent's actions as randomized interventions and uses per-dimension two-sample tests with FDR correction to identify controllable observation dimensions, matching oracle returns across 12 continuous-control tasks with up to 100 distractors.

Jiaxin Liu, Anzhe Cheng, Paul Bogdan

Sydney Poster Session 4, Wed, Dec 9, 5:00 PM–8:00 PM, Hall 1-4 · Published 2026

Dual-Pathway Circuits of Object Hallucination in Vision-Language Models

Vision-language models contain separate visual grounding and hallucination pathways whose components flip polarity to drive errors, and suppressing them cuts object hallucination by up to 76%.

Jiaxin Liu, Ding Zhong, Yue Wang, Zhidong Yang and 5 more

Sydney Poster Session 1, Tue, Dec 8, 10:00 AM–1:00 PM, Hall 1-4 · Published 2026

Instructing LLMs to Negotiate using Reinforcement Learning with Verifiable Rewards

RLVR trains a 30B LLM buyer via verifiable economic rewards to negotiate, revealing four-phase strategic evolution and outperforming much larger frontier models in surplus extraction.

Shuze D Liu, Claire Chen, Jiabao S Xiao, Lei Lei and 3 more

Atlanta Poster Session 5, Fri, Dec 11, 10:00 AM–1:00 PM, Hall C1 · Published 2026

Tiny but Trusted: Efficient Vision-Language Reasoning for Time-Series Anomaly Detection

VisAnomReasoner is a parameter-efficient vision-language model that improves time-series anomaly detection precision and F1 by over 21 points via VisAnomBench fine-tuning with natural-language rationales.

Xiaona Zhou, Tianjiao (Joey) Yu, Muntasir Wahed, Constantin Brif and 1 more

Atlanta Poster Session 4, Thu, Dec 10, 4:30 PM–7:30 PM, Hall C1 · Published 2026 · ▲ 2 on Hugging Face

Decoding the Critique Mechanism in Large Reasoning Models

Large reasoning models use hidden critique abilities to recover from uncorrected reasoning errors, and steering with critique vectors improves error detection and test-time scaling without training.

Hoang Phan, Nguyen Hung-Quang, Thanh Quoc Hung Le, Xiusi Chen and 2 more

Sydney Poster Session 5, Thu, Dec 10, 10:00 AM–1:00 PM, Hall 1-4 · Published 2026 · Code ★ 1

LangFlow: Continuous Diffusion Rivals Discrete in Language Modeling

LangFlow closes the continuous-discrete language-modeling gap via flow matching and a learnable noise schedule, matching discrete diffusion perplexity and exceeding autoregressive zero-shot results on four benchmarks.

Yuxin Chen, Chumeng Liang, Hangke Sui, Ruihan Guo and 3 more

Sydney Poster Session 6, Thu, Dec 10, 5:00 PM–8:00 PM, Hall 1-4 · Published 2026 · ▲ 15 on Hugging Face · Code ★ 96

OpenClawBench: Benchmarking Process-side Anomalies in Real-world Agent Execution Trajectories

OpenClawBench benchmarks process-side agent anomalies via 31,264 annotated trajectories, revealing 2,904 process failures among 31,135 oracle-passing executions.

Yibing Liu, Yangze Liu, Xiao-Long Yin, Bin Wang and 3 more

Sydney Poster Session 2, Tue, Dec 8, 5:00 PM–8:00 PM, Hall 1-4 · Published 2026

LESSViT: Robust Hyperspectral Representation Learning under Spectral Configuration Shift

LESSViT enables robust cross-sensor hyperspectral representation via low-rank spatial-spectral attention, channel-agnostic embeddings, and masked pretraining with decoupled masking and hierarchical sampling.

Haozhe Si, Yuxuan Wan, Yuqing Wang, Minh Do and 1 more

Atlanta Poster Session 1, Wed, Dec 9, 10:00 AM–1:00 PM, Hall C1 · Published 2026

TriAxialKV: Toward Extreme Low-Precision KV-Cache Quantization for Agentic Inference Tasks

TriAxialKV assigns triaxial tags to KV-cache tokens and uses per-tag sensitivity to allocate INT2/INT4 under fixed memory, matching BF16 accuracy with 4.5x cache compression and 30% higher throughput on agentic tasks.

Hanzhang Shen, Haoran Wu, Yiren Zhao, Robert Mullins

Sydney Poster Session 3, Wed, Dec 9, 10:00 AM–1:00 PM, Hall 1-4 · Published 2026