HuggingFace

Articles from HuggingFace

Cards List

HypoEvolve: Genetic Algorithms Enable Multi-Agent LLMs to Discover Scientific Hypotheses

Hugging Face Daily Papers ↗ · 2026-09-14 Cached

HypoEvolve introduces a framework using genetic algorithms to coordinate multi-agent LLMs for scientific hypothesis discovery, demonstrated through drug repurposing in cancer research with improved performance over baselines.

0 favorites 0 likes

HarnessVLN: Unifying Training-Free Embodied Navigation through an Agent Harness

Hugging Face Daily Papers ↗ · 2026-09-14 Cached

HarnessVLN is a zero-shot, training-free framework for embodied navigation that unifies perception, retrieval, grounding, navigation, recovery, and termination through a unified tool interface, achieving state-of-the-art results on benchmarks like R2R and RxR.

0 favorites 0 likes

ModularRSI: Modular and Generalizable Recursive Harness Self-Improvement

Hugging Face Daily Papers ↗ · 2026-09-14 Cached

ModularRSI introduces a modular and generalizable framework for recursive self-improvement in AI agent harnesses, using contrastive learning across tasks to evolve modules independently and enhance performance on unseen tasks.

0 favorites 0 likes

The Router Within: Eliciting Native Skill Routing from a Frozen LLM

Hugging Face Daily Papers ↗ · 2026-09-14 Cached

The paper introduces Gavel, a method that elicits native skill routing from frozen LLMs via linear projections, enabling efficient tool selection without context overload and outperforming existing pipelines on benchmarks.

0 favorites 0 likes

Disentangling Representation Evolution in Transformers through Directional Decomposition

Hugging Face Daily Papers ↗ · 2026-09-14 Cached

This paper decomposes transformer representation updates into parallel and perpendicular components to study evolution geometry, linking it to editing robustness, compression diagnosis, and training improvements.

0 favorites 0 likes

ModaLens: Measuring Image Sensitivity in Report-Conditioned Medical VLMs

Hugging Face Daily Papers ↗ · 2026-09-14 Cached

ModaLens is a paired image-swap audit that measures how report availability reduces image sensitivity in medical vision-language models, demonstrated using MedGemma-27B on the MIMIC-CXR dataset.

0 favorites 0 likes

How Lossless Is Lossless Speculative Decoding? The Role of Numerical Precision in Orthrus

Hugging Face Daily Papers ↗ · 2026-09-14 Cached

This paper investigates the losslessness of Orthrus, a hybrid autoregressive-diffusion model for inference acceleration, finding that it requires high numerical precision (FP32) for exact trajectory matching, while BF16 divergence does not impair downstream performance.

0 favorites 0 likes

LynnReal-Omni: Native multi-modal Video Generation for Agentic Visual Workflows

Hugging Face Daily Papers ↗ · 2026-09-14 Cached

LynnReal-Omni is a unified multimodal video diffusion framework that integrates agentic visual controls with high-fidelity generation and real-time acceleration for stable, controllable video creation.

0 favorites 0 likes

Not All Prompts Are Equal: Exploration-Guided Prompt Scaffolding for Multimodal Reinforcement Post-Training

Hugging Face Daily Papers ↗ · 2026-09-14 Cached

This paper proposes an exploration-guided prompt scaffolding framework that dynamically adjusts training prompts for multimodal reinforcement learning, achieving up to 9.7% relative improvement in performance on benchmarks.

0 favorites 0 likes

PhysBrain 1.5: From Vision-Language Models to Physical Foundation Models

Hugging Face Daily Papers ↗ · 2026-09-14 Cached

PhysBrain 1.5 is a unified model that integrates physical environment understanding, action generation, and future state prediction via autoregressive training, achieving state-of-the-art open-source performance on 28 embodied benchmarks.

0 favorites 0 likes

Kaininja: Extending Native 3D Generators to the Part Level

Hugging Face Daily Papers ↗ · 2026-09-14 Cached

KaiNinja extends a native 3D generator to produce part-level outputs using a dual-volume representation, improving both part and whole-object fidelity without requiring segmentation.

0 favorites 0 likes

Atria Dawn: The Dawn of Agentic Superintelligence

Hugging Face Daily Papers ↗ · 2026-09-14 Cached

Atria Dawn Preview is a foundation agentic language model designed for scientific research, achieving competitive benchmark results and demonstrating a shift toward human-AI project-level collaboration.

0 favorites 0 likes

When Agents Slow Down: Understanding LLM Agents' Test-Time Strategies via Elo-per-token Analysis

Hugging Face Daily Papers ↗ · 2026-09-14 Cached

This paper introduces Elo-per-token analysis to study how LLM agents allocate test-time compute, revealing that agents initially outperform independent sampling but slow down over time, with parallel sessions offering performance gains.

0 favorites 0 likes

Enabling Creative Exploration for Vibe Design Agents

Hugging Face Daily Papers ↗ · 2026-09-14 Cached

This paper introduces a method for vibe design agents to explore diverse UI alternatives by separating exploration from implementation through structured design specifications, as evaluated on 168 prompts and a large online experiment with over 300,000 tasks.

0 favorites 0 likes

Pick Your Poison: Learning to Select Poison Sets for Stronger LLM Backdoor Attacks

Hugging Face Daily Papers ↗ · 2026-09-14 Cached

This paper introduces SAILS, a method for selecting optimal poison sets in backdoor attacks against large language models, improving worst-case attack success by 30 percentage points over baselines.

0 favorites 0 likes

HazardAuditor: From Executable Threats to Safer Computer-Use Agents

Hugging Face Daily Papers ↗ · 2026-09-14 Cached

HazardAuditor introduces an execution-grounded framework for supervising safety in computer-use agents, with Guard Policy Optimization improving safety outcomes by up to 16.5% accuracy over prior methods.

0 favorites 0 likes

RSIAgent: Autonomous Exploration for Recursive Self-improvement in New Environments

Hugging Face Daily Papers ↗ · 2026-09-14 Cached

RSIAgent is a training-free multi-agent framework that enables digital agents to adapt to new environments through recursive self-improvement, autonomous memory construction, and broad-then-deep exploration, outperforming closed-source models on benchmarks.

0 favorites 0 likes

Omni-Streaming Thinking

Hugging Face Daily Papers ↗ · 2026-09-14 Cached

Omni-Streaming Thinking improves streaming omni-modal reasoning by deferring claims until cross-modal verification, reducing premature commitment and auditory hallucinations.

0 favorites 0 likes

Dream-RSI: Recursive Self-Improvement through Evolving Worlds

Hugging Face Daily Papers ↗ · 2026-09-14 Cached

Dream-RSI is a framework for scalable recursive self-improvement in AI agents that uses historical discovery trees to create a replay simulator for offline policy evaluation, reducing online costs and improving discovery efficiency.

0 favorites 0 likes

BVB: Benchmarking Agentic Video Understanding via Programmatic Reconstruction in Blender

Hugging Face Daily Papers ↗ · 2026-09-14 Cached

The paper introduces BVB, a benchmark for agentic video understanding via programmatic reconstruction in Blender, evaluating models on perceptual similarity and spatiotemporal fact retention.

0 favorites 0 likes
← Previous
Next →
← Back to home

Submit Feedback