## Episode Summary

In this episode, we cover:

- **Model or Harness? An Interaction-Centric Taxonomy for Localizing Agent Failures** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2607.28802)

- **GPTQ-2D: Cubic-Time Two-Sided Adaptive Rounding** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2607.27042)

- **Seeing or Knowing? Visual Context Sensitivity in Multimodal Large Language Models** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2607.26326)

- **Zero-Mem: Zero-Token Memory Operations for LLM Agents** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2607.29377)

- **RecHarness: A Bandit-Routed Agentic Harness for Self-Evolving Recommender Systems** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2607.29241)

---

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

## Episode Summary

In this episode, we cover:

- **SAF-OPD: Stable Advantage Fusion for On-Policy Distillation** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2607.29209)

- **EMBL AI Librarian: Life-Sciences Knowledge Layer for AI Agents** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2607.28229)

- **Would You Walk to the Car Wash? Revealing the Salience Bias of Large Language Models in Commonsense Reasoning** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2607.28478)

- **CodeShrink: Adaptive Visual Compression for Efficient Multimodal Code Understanding** (arXiv)

- [Read more](http://arxiv.org/abs/2607.29637v1)

- **Fewer Clarifications, Better Code: Benchmarking Cross-Session Personalized Ambiguity Adaptation in Coding Assistants** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2607.26611)

---

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

## Episode Summary

In this episode, we cover:

- **Fairness Pruning: Locating Demographic Bias in GLU-MLP Layers via Differential Activations** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2607.28319)

- **Explorative Modeling: Unlocking a Third Pretraining Axis and End-to-End Generation** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2607.27372)

- **Filesystem-Based Memory for LLM Agents: Organization, Evolution, and Sustainability** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2607.26637)

- **AskChem: Claim-Centered Infrastructure for Chemistry Literature Synthesis** (arXiv)

- [Read more](http://arxiv.org/abs/2607.28618v1)

- **Σ-Mem: An Online Reliability Memory for LLM-based Multi-Agent Systems** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2607.27958)

---

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

## Episode Summary

In this episode, we cover:

- **Multi-Head Attention Residuals** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2607.27230)

- **PAIChecker: Uncovering and Checking PR-Issue Misalignment in SWE-Bench-Like Benchmarks** (arXiv)

- [Read more](http://arxiv.org/abs/2607.28587v1)

- **Beacon: Knowing When and How to Perform Agentic Visual Reasoning** (arXiv)

- [Read more](http://arxiv.org/abs/2607.28595v1)

- **β-OPSD: Deriving with Policy Optimization, Training with Self-Distillation** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2607.28582)

- **ShadowDancer: Teaching Video World Models Any Action by Learning Unified Dynamics Representations from a Video and Its Shadow** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2607.28362)

---

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

## Episode Summary

In this episode, we cover:

- **OSReward: Instituting Standardized Evaluation for Cross-Platform Computer-Use Reward Models** (arXiv)

- [Read more](http://arxiv.org/abs/2607.28609v1)

- **Change2Task: From Repository Changes to Executable Coding Agent Tasks and Environments** (arXiv)

- [Read more](http://arxiv.org/abs/2607.28591v1)

- **LEDGERMIND: Provenance-Constrained Multimodal Agentic Reasoning with a Structured Evidence Ledger** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2607.28374)

- **See2Think: Do Multimodal Models Really Use Intermediate Visual States?** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2607.26769)

- **Chimera: Designing and Chinchilla-Scaling Hybrid Visual Diffusion Transformers** (arXiv)

- [Read more](http://arxiv.org/abs/2607.28611v1)

---

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

## Episode Summary

In this episode, we cover:

- **SpecFirst: Behavioral Specification Elicitation as a First-Class Step in Agent-Based Program Synthesis from Scratch** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2607.27167)

- **MindForge: Teaching Small Language Models Whole-Life-Cycle Software Engineering via Source-Free Program Synthesis** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2607.27146)

- **Voice Memory for Agentic Speech Recognition** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2607.26410)

- **CLBench-V: Evaluating Multimodal Context Learning from Grounding to Knowledge Acquisition** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2607.25294)

- **CAST: Game Solvers as Turn-Level Teachers for LLM Agents** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2607.25308)

---

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

## Episode Summary

In this episode, we cover:

- **MODUS: Decoder-Only Any-to-Any Modeling of Diverse Modalities** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2607.25948)

- **VetClaw: An Edge-Cloud Multimodal Agentic System for Veterinary Disease Screening** (arXiv)

- [Read more](http://arxiv.org/abs/2607.26042v1)

- **Beyond Zooming: Learning Multi-Tool Visual Reasoning for Ultra-High-Resolution Remote Sensing** (arXiv)

- [Read more](http://arxiv.org/abs/2607.25993v1)

- **Parallel Decoding Distillation for Fast Image and Video Generation** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2607.26004)

- **MDTransformer: A Hardware-Software Co-Design of Mode-Division Photonic Transformer Accelerator with Inverse-Designed Coherent Crossbar** (arXiv)

- [Read more](http://arxiv.org/abs/2607.26016v1)

---

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

## Episode Summary

In this episode, we cover:

- **TILT: Improving Compositional Generation in Diffusion Models with a Model-Intrinsic Reward** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2607.21606)

- **A Vocabulary for Multi-Agent Automated Research Systems** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2607.22682)

- **Rethinking Classifier-Free Guidance in On-Policy Diffusion Distillation** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2607.24731)

- **The Physics of Multi-Turn Long-Horizon Planning: From Pre-training to Post-training via Single- and Multi-Teacher On-Policy Agentic Distillation** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2607.24720)

- **Sol-Attn: Accelerating Video Generation Inference via On-the-Fly Attention Sparsification** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2607.24027)

---

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

## Episode Summary

In this episode, we cover:

- **Scaling Native Multimodal Pre-Training From Scratch** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2607.22043)

- **DataPrep-Bench: Benchmarking LLMs as Training Data Preparators** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2607.20465)

- **Skill Self-Play: Pushing the Frontier of LLM Capability with Co-Evolving Skills** (arXiv)

- [Read more](http://arxiv.org/abs/2607.22529v1)

- **Opaque Epistemic Mediation: How LLM Deployment Configurations Shape the Validation of Pseudo-Science** (arXiv)

- [Read more](http://arxiv.org/abs/2607.22513v1)

- **Molt: A Scalable PyTorch-Native Training Framework for Agentic Reinforcement Learning** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2607.21653)

---

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

## Episode Summary

In this episode, we cover:

- **FinanceComplexQA: Benchmarking Agentic Reasoning on Industrial-grade Financial Documents** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2607.19238)

- **MedGame: Storytelling Gamification Empowered by Large Language Models for Medical Education** (arXiv)

- [Read more](http://arxiv.org/abs/2607.21570v1)

- **Beyond Sycophancy: Structured Resistance and Compliance in LLM Moral Reasoning** (arXiv)

- [Read more](http://arxiv.org/abs/2607.21558v1)

- **3D-Aware VLMs with Implicit and Explicit Geometries** (arXiv)

- [Read more](http://arxiv.org/abs/2607.21595v1)

- **K12-KGraph: A Curriculum-Aligned Knowledge Graph for Benchmarking and Training Educational LLMs** (Hugging Face Daily)

- [Read more](https://huggingface.co/papers/2605.09635)

---

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