Unzip

## Episode Summary

In this episode, we cover:

- **Cross-Prompt Reuse of Entity-Replacement Responses** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2611.00352)

- **Kierston & KierstonBench: Simulation of Allentown, PA and Evaluation of Language Models in Crime Investigation and Surveillance Tasks** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2611.04785)

- **Adversarial Speculative Reasoning Improves Long Horizon Agents** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2611.05345)

- **SphereFlow: Missing Modality Imputation via Geometric Transport on Hypersphere** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2611.04414)

- **When Fine-Grained Credit Assignment Helps and Hurts: Multimodal Agentic Reinforcement Learning through Video Editing** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2610.35497)

---

*Sponsored by Ada, Ago Consulting, and Zen Semiconductor*

## Episode Summary

In this episode, we cover:

- **Metacognitive Decoding: Efficient Inference for Diffusion Language Agents** (arXiv)

- [Read more](http://arxiv.org/abs/2610.37642v1)

- **TAO: Tool-Aware Joint Scheduling for LLM-Based Agentic Serving** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2610.37230)

- **Efficient Streaming Audio-Visual Target Speaker Extraction for Real-World Acoustic Scenes** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2610.32749)

- **CausGround: Causal Interaction Grounding for Adaptive Embodied Planning** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2610.29702)

- **Amortizing Sequential Experimental Design in Reasoning LLMs** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2610.20832)

---

*Sponsored by Ada, Ago Consulting, and Zen Semiconductor*

## Episode Summary

In this episode, we cover:

- **h1: A Taste of the Bitter Lesson in Automated Theorem Proving** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2610.35962)

- **Multiscale Policy Optimization for Long-Horizon LLM Agents** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2610.35013)

- **GEPA-Curator: Curated Learning for Self Improving Agents** (arXiv)

- [Read more](http://arxiv.org/abs/2610.37651v1)

- **Heads of Security: Learned Attention Gates Improve Prompt Injection Robustness** (arXiv)

- [Read more](http://arxiv.org/abs/2610.37657v1)

- **AlphaBet: Reinforcement Learning for LLM Event Forecasting Agents** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2610.22684)

---

*Sponsored by Ada, Ago Consulting, and Zen Semiconductor*

## Episode Summary

In this episode, we cover:

- **Reva: Real-Time Vulnerability Detection through Contextual Probing of Large Language Models** (arXiv)

- [Read more](http://arxiv.org/abs/2610.35829v1)

- **Profiling a Frontier Agent with EBench: From Direct Control to Harness with Policy** (arXiv)

- [Read more](http://arxiv.org/abs/2610.35832v1)

- **Understanding Jev-Style System One Models: Does Structure Reduce Selection Bias?** (arXiv)

- [Read more](http://arxiv.org/abs/2610.35826v1)

- **CRM-Agent-Bench: Benchmarking Tool-Using Agents in Stateful Enterprise Environments** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2610.32004)

- **EnvLunch: The Environment as a Free Lunch for Learning to Interact** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2610.33817)

---

*Sponsored by Ada, Ago Consulting, and Zen Semiconductor*

## Episode Summary

In this episode, we cover:

- **Kolab: Benchmarking Human-Agent Collaboration in Long-Horizon Workflows** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2610.21109)

- **On the Token Collapse Phenomenon in Textual Graph Soft Prompting** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2610.31572)

- **Reading, Writing, Sharing, and Interacting with the World through Video** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2610.31668)

- **JockeyBench: Evaluating LLM Agents on Horse-Racing Decision-Making** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2610.19589)

- **Speculative Diffusion is a Tale of Two Tradeoffs** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2610.31819)

---

*Sponsored by Ada, Ago Consulting, and Zen Semiconductor*

## Episode Summary

In this episode, we cover:

- **Direct Pass@$k$ Optimization for Thought Pruning: Balancing Quality and Diversity in Test-Time Reasoning** (arXiv)

- [Read more](http://arxiv.org/abs/2610.31667v1)

- **Different Tasks, Different Timesteps: Temporal Task-Aware Quantization for Diffusion Language Models** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2610.31198)

- **Agent Mosaic: Recognizing Confidentiality Leaks in Agentic LLM Systems** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2610.30418)

- **Realism VS Accuracy: Event Sequence Forecasting from a Generative Modeling Perspective** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2610.12586)

- **BrickWright: Agentic Creation of Physically Feasible Brick Builds** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2610.31651)

---

*Sponsored by Ada, Ago Consulting, and Zen Semiconductor*

## Episode Summary

In this episode, we cover:

- **The LLM-Attacker: A Unified Red-Team Framework for Privacy Evaluation of Split and Distributed LLM Systems** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2610.22722)

- **Can an Agent Debug a Web Corpus? Agentic Auditing of Slovak FineWeb2** (arXiv)

- [Read more](http://arxiv.org/abs/2610.31636v1)

- **Bridging Exploration and Persona: Information-Directed Dueling for Test-Time Agent Personalization** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2610.24385)

- **EvalResearchBench: Can AI Agents Build Their Own Feedback for Recursive Self-Improvement?** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2610.04184)

- **Who Takes the Lead? Model Ability and Task-Conditioned Collaboration with Humans** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2610.27148)

---

*Sponsored by Ada, Ago Consulting, and Zen Semiconductor*

## Episode Summary

In this episode, we cover:

- **IF-MAP: Influence-Function-Inspired Agent Prioritization in Multi-Agent Systems** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2610.24383)

- **Diffusion Language Models Change Speed Across Tasks and Languages** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2610.25835)

- **EpiSEM-Bench: Evaluating Safety and Evidence Recovery in LLM Agents under Missing Evidence** (arXiv)

- [Read more](http://arxiv.org/abs/2610.31650v1)

- **FilmFlow: Agentic System for 3D Virtual Film Cinematography** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2610.30753)

- **METTA: Memory-Enhanced Test-Time Adaptation for Meta-Reinforcement Learning** (arXiv)

- [Read more](http://arxiv.org/abs/2610.31656v1)

---

*Sponsored by Ada, Ago Consulting, and Zen Semiconductor*

## Episode Summary

In this episode, we cover:

- **ToM2Vec: Continuous Theory-of-Mind Tokens for Multi-Agent Coordination** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2610.29934)

- **DiffResearch: Accelerating Deep Research with Diffusion Language Models** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2610.28909)

- **Looped Uniform Diffusion Language Models** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2610.30066)

- **Did the Problem Actually Get Harder? Testing LLMs on Paired P and NP-Complete Problems Hardened by Adversarial Multi-Agent Rewriting** (arXiv)

- [Read more](http://arxiv.org/abs/2610.30103v1)

- **FrontierCS 2: A Benchmark for Computer Science R&D** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2610.29932)

---

*Sponsored by Ada, Ago Consulting, and Zen Semiconductor*

## Episode Summary

In this episode, we cover:

- **Training the Whole Agent, Not Just the Model** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2610.28069)

- **Self-Credit: Credit Assignment for Long-Horizon Agentic RL via Intra-group Rescaling** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2610.28388)

- **On the Tip of the Tongue: First-Token Cueing as a Behavioral Probe of Recall Failures in LLMs** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2610.23400)

- **EAU: Environment-Agent-User Co-evolution for Training User-Centric Agents** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2610.16523)

- **Axiomatic World Modeling for Physics Reasoning** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2610.24519)

---

*Sponsored by Ada, Ago Consulting, and Zen Semiconductor*