Daily AI Papers — August 10, 2026

1. SFT Conflicts, RL Coexists: A Theoretical and Empirical Analysis of Multi-Task Learning for LLMs

Authors: Kejian Zhu, Zhuoran Jin, Shangqing Tu, Hongbang Yuan, Yushi Bai, Kang Liu, Juanzi Li, Jun Zhao arXiv: arxiv.org/abs/2608.03573 Summary: Supervised Fine-Tuning (SFT) and Reinforcement Learning (RL) exhibit fundamentally different behaviors in enhancing multi-task reasoning for large language models (LLMs). Our preliminary experiments revealed a phenomenon: SFT suffers from severe task conflicts under multi-stage training, whereas RL enables stable coexistence across diverse tasks. Trending because: 29 HuggingFace upvotes + one of the most-upvoted papers in today’s feed.

2. Beyond Simply Environment Scaling: Designing Effective Environment Distributions for Multimodal Agent Learning

Authors: Kejian Zhu, Zhuoran Jin, Dongqi Huang, Hongbang Yuan, Yupu Hao, Kang Liu, Jun Zhao arXiv: arxiv.org/abs/2608.03571 Summary: Recent works train agents by constructing large-scale multimodal environment pools. However, we find that simply increasing the number of multimodal environments does not always benefit. Trending because: 28 HuggingFace upvotes + one of the most-upvoted papers in today’s feed.

3. SimWAM: A Simple World Action Model for End-to-End Autonomous Driving

Authors: Zongchuang Zhao, Xin Zhou, Tianyang Xu, Zhengyang Sun, Kaixuan Zhou, Honglin Li, Dingkang Liang, Xiang Bai arXiv: arxiv.org/abs/2608.07468 Summary: World-Action Models (WAMs) improve end-to-end autonomous driving by transferring video dynamics priors to action prediction, but existing methods require costly future generation at inference. We present SimWAM, a simple yet effective WAM that uses video generation purely as a training signal. Trending because: 22 HuggingFace upvotes + one of the most-upvoted papers in today’s feed.

4. StreamArena: Toward Continuous, Interactive, and Long-Horizon Agentic Streaming Video Understanding

Authors: Xichen Zhang, Guankai Li, Yinghao Zhu, Shijian Wang, Sitong Wu, Shaozuo Yu, Meng Chu, Yuan Lu, Jiaya Jia arXiv: arxiv.org/abs/2608.05703 Summary: Deploying autonomous multimodal agents in continuous, real-world environments requires them to ingest unbounded audio-visual streams and maintain hour-scale memory. However, current evaluations predominantly rely on brief clips and multiple-choice formats. Trending because: 14 HuggingFace upvotes + high engagement among today’s fresh releases.

5. YOLO-PEFT: Parameter-Efficient Fine-Tuning on YOLO Family

Authors: Xu Lin, WenJie Nie, Jinlong Peng, Weifu Fu, YueXiao Ma, Xiawu Zheng, Yong Liu arXiv: arxiv.org/abs/2608.07051 Summary: Generic parameter-efficient fine-tuning (PEFT) methods transferred from language models can fail silently on real-time detectors, whose heterogeneous operators and detection-specific components impose placement constraints absent from regular Transformer stacks. We propose YOLO-PEFT, a structure-aware framework that formulates adapter placement as an auditable constraint-planning problem. Trending because: 14 HuggingFace upvotes + high engagement among today’s fresh releases.

6. When Activation Oracles Learn Not to Read: Concept-Specific Blind Spots in Fine-Tuned Oracles

Authors: Tobias Bersia, Tatiana Gaintseva arXiv: arxiv.org/abs/2607.23379 Summary: Activation Oracles (AOs) are language models trained to answer natural-language questions about another model’s internal activations. They offer a flexible interface for reading hidden information from model states, especially when relevant information is internally represented but absent or incomplete in visible behavior. Trending because: 10 HuggingFace upvotes + high engagement among today’s fresh releases.

7. Efficient Knowledge Distillation for LLMs: Offline Top-K Logits and a Fused Chunked KL Loss

Authors: Bakbergen Ryskulov, Iker García-Ferrero, David Montero, David Jansen, Ali Hashemi, Jezabel R. Garcia, Antonio Tiene, Román Orús arXiv: arxiv.org/abs/2608.03796 Summary: Small language models are often the only option for deployment under tight latency, cost, and on-premises constraints, but they are rarely trained from scratch: a compressed model is usually recovered through knowledge distillation (KD). This recovery step largely decides the final quality, yet it is expensive. Trending because: 9 HuggingFace upvotes + high engagement among today’s fresh releases.

8. Reinforcement Learning with Evolving Rubrics as Rewards for Audio Reasoning

Authors: Fangxu Yu, Tao Feng, Dehai Min, Zinan Lin, Weijia Xu, Michael Xu, Philip S. Yu, Ge Liu, Tianyi Zhou arXiv: arxiv.org/abs/2608.02831 Summary: Audio reasoning is essential for machine understanding of the acoustic world. Reinforcement learning with verifiable rewards can elicit such reasoning, yet existing reward designs are complementary in their limitations: outcome-based rewards supervise only the final answer and let the model reach it without attending to the audio, whereas process-based rewards score the reasoning itself but rely on coarse, hand-crafted, and fixed criteria that neither adapt to each question nor stay grounded in the acoustic evidence. Trending because: 9 HuggingFace upvotes + high engagement among today’s fresh releases.

9. Douyin Multimodal Embedding Model Technical Report

Authors: Haonan Chen, Chu Li, Zhicheng Wang, Yuanwei Liu, Yuanjiang Wang, Shaohua Jiang, Zhicheng Dou arXiv: arxiv.org/abs/2608.02148 Summary: Multimodal representation learning is a cornerstone of modern AI. By encoding multimodal queries and targets into vectors, it powers industrial search and recommendation and underpins modern agents. Trending because: 7 HuggingFace upvotes + solid early traction in today’s feed.

10. Relevant but Incomplete: Referential Dangling as a Paradigm-Level Failure Mode in Hard Prompt Compression

Authors: Zhengpei Hu, Kai Li, Dapeng Fu, Xuechao Zou, Yuanhao Tang, Yue Li, Tengfei Cao, Jianqiang Huang arXiv: arxiv.org/abs/2608.04569 Summary: Hard prompt compression reduces long-context inference cost by independently scoring tokens, sentences, or chunks and retaining the highest-scoring units under a budget. We identify a structural failure in this procedure: independent selection can split dependent evidence pairs, retaining one member while deleting the other. Trending because: 7 HuggingFace upvotes + solid early traction in today’s feed.

11. Uncertainty-Aware World Model for Aerial Image-Goal Navigation

Authors: Deyi Zhu, Haoyu Fan, Yinan Zhu, Weichen Zhang, Shilin Ma, Xinlei Chen, Yansong Tang arXiv: arxiv.org/abs/2608.05597 Summary: Aerial image-goal navigation requires an unmanned aerial vehicle (UAV) to reach a target location specified by a goal image. Existing world-model-based methods rank candidate trajectories using predicted futures, but typically rely on only one or a few point predictions, which is inadequate for large-scale outdoor environments with substantial future-state uncertainty. Trending because: 7 HuggingFace upvotes + solid early traction in today’s feed.

12. Addressable Memory for Video World Models

Authors: Xindi Wu, Sven Elflein, James Lucas, Olga Russakovsky, Laura Leal-Taixé, Despoina Paschalidou, Jonathan Lorraine, Aljoša Ošep arXiv: arxiv.org/abs/2608.07408 Summary: We study visual persistence in interactive video world models. These models rely on a Key-Value (KV) cache as a growing visual memory to carry forward previously generated frames. Trending because: 6 HuggingFace upvotes + solid early traction in today’s feed.

13. Round-Trip Consistency: Bidirectional Diffusion Models Can Predict Their Own Rollout Errors

Authors: Alexander Scheinker arXiv: arxiv.org/abs/2608.00675 Summary: Autoregressive models accumulate error over long rollouts, yet at deployment there is no ground truth to measure it against. We train a single conditional latent diffusion model that steps a dynamical system forward or backward in time via a direction flag, and show that this bidirectionality supplies a measurement-free test-time error signal: rolling forward i steps and then backward i steps must return the model to its start, so the round-trip discrepancy C_i is a self-supervised proxy for the unobservable rollout error: no ensembles, no held-out data, no governing equations, for one extra rollout. Trending because: 6 HuggingFace upvotes + solid early traction in today’s feed.

14. Modular TTT: Rethinking Test-Time Training as Composable Modules

Authors: Bohao Tang, Zhen Qin, Yuqi Pan, Zheng Li, Pengfei Liu, Ya Zhang arXiv: arxiv.org/abs/2608.07110 Summary: Test-time training (TTT) views sequence modeling as an online learning problem in which fast weights are updated by an internal learning rule. Despite the growing number of TTT variants, existing approaches typically hard-code each variant separately, which makes it difficult to design new TTT methods and to isolate the role of each component. Trending because: 5 HuggingFace upvotes + solid early traction in today’s feed.

15. Skaling: Chinchilla’s Exponents Meet Kaplan’s Coupling

Authors: Mathurin Videau, Badr Youbi-Idrissi, David Lopez-Paz, Kartik Ahuja arXiv: arxiv.org/abs/2608.07222 Summary: Neural scaling laws are foundational for language model development, yet standard formulations systematically under- and overestimate loss at data-scarce and overtraining extremes. This failure originates in the underlying assumption that model size and training data impact the loss independently. Trending because: 5 HuggingFace upvotes + solid early traction in today’s feed.

16. Characterizing the Quality Profile of AI-Generated C++ in Production

Authors: Michael Tran, Fred Lewis, Kun Yang, Saksham Thakur, Aditya Kini, Aditya Patil, Milad Hashemi, Parthasarathy Ranganathan arXiv: arxiv.org/abs/2608.06640 Summary: The widespread integration of AI coding assistants offers undeniable boosts to engineering velocity. Yet, recent studies point to a growing trade-off, revealing persistent challenges with code quality and maintainability. Trending because: 4 HuggingFace upvotes + solid early traction in today’s feed.

17. Small Foundation Models of Human Cognition and Behaviour

Authors: Nick Oh, Fernand Gobet arXiv: arxiv.org/abs/2608.05224 Summary: Large language models fine-tuned on human behavioural data have emerged as general-purpose cognitive proxies, but the scale this requires, and whether these models process task structure or exploit statistical shortcuts, remain open questions. We train fourteen models from 135M to 14B parameters across four architecture families on Psych-101, a dataset of 10.7 million trial-level choices from 160 experiments. Trending because: 4 HuggingFace upvotes + solid early traction in today’s feed.

18. The Optimizer Is the Agent: Reasoning-Driven Search across Prompts, Programs, and ML Workflows

Authors: Junbo Li, Boyi Liu, Canwen Xu, Yite Wang, Yuxiong He, Zhangyang Wang, Qiang Liu, Zhewei Yao arXiv: arxiv.org/abs/2608.06714 Summary: Recent systems for optimizing prompts, programs, and ML workflows typically rely on explicit outer-loop controllers such as evolutionary search, bandits, or textual-gradient methods. We ask a fundamentally different question: how much of this search policy can be internalized by a single tool-using agent? Trending because: 4 HuggingFace upvotes + solid early traction in today’s feed.

19. Do AI Personas Grow? Analyzing and Benchmarking Personality Evolution in LLM Agents After Life Events

Authors: Ming Wang, Peidong Wang, Xiaocui Yang, Daling Wang, Shi Feng, Fiona Fui-Hoon Nah, Ee-Peng Lim arXiv: arxiv.org/abs/2608.06485 Summary: Personality-conditioned LLM agents (PC-Agents) are increasingly used in emotional support, social simulation, and role-playing, motivating the development of lifelong agents that remain coherent over extended interactions. A key component of such coherence is personality evolution: agents should undergo plausible, psychology-grounded changes as they experience life events in different contexts. Trending because: 3 HuggingFace upvotes + solid early traction in today’s feed.

20. When Privileged Guidance Misaligns: State-Matched Routing and Contextualized Self-Distillation for Multi-Turn Agents

Authors: Junzhuo Liu, Weiwei Li, Jun Ling, Peng Wang arXiv: arxiv.org/abs/2608.05219 Summary: Privileged on-policy distillation provides dense supervision for multi-turn agents by allowing a synchronized teacher to re-score the student’s response at every turn with access to training-only references, such as successful trajectories. In interactive environments, however, the student’s preceding actions continually change the execution state. Trending because: 3 HuggingFace upvotes + solid early traction in today’s feed.