Search: “gpt”

Normalize-Then-Precondition: A Hierarchical Approach to Marginal Scale and Interaction Geometry for LLM Training

arXiv:2609.36692v1 Announce Type: cross Abstract: Matrix optimizers have emerged as a promising direction, with Muon standing out as a prominent design. Revisiting Muon through its full-Gram represent

Content type

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

Industry

OpenAI says individuals associated with Moonshot AI played a significant role in a coordinated model-distillation campaign that began in early July (Maggie Eastland/Bloomberg)

DGX agent

Maggie Eastland / Bloomberg: OpenAI says individuals associated with Moonshot AI played a significant role in a coordinated model-distillation campaign that began in early July — OpenAI accused its Ch

industrytechmeme

30 Sept 2026

RareDx: Controlled Knowledge Integration and Graph-Grounded Policy Optimization for Rare-Disease Diagnosis

arXiv:2609.35549v2 Announce Type: replace Abstract: Rare-disease diagnosis is a long-tail reasoning problem: phenotypes are incomplete, individual disorders are sparsely documented, and relevant evide

Safety

Real2Gym: Building Gyms from Videos, Bringing Skills to Robots

arXiv:2609.37089v1 Announce Type: new Abstract: Real-world videos provide rich demonstrations of manipulation, but turning them into reusable robot skills requires visually aligned environments, execu

safetyarxiv-cs-cv

Representational Simplicity and Circuit Size Dissociate in a Threshold-Dependent Way: A Controlled Test via Adversarial Training

arXiv:2609.35890v1 Announce Type: new Abstract: Sparse-autoencoder decomposability and concentrated feature attribution are increasingly treated as evidence that a model's computation is easier to rev

researcharxiv-cs-ai

RoboChrono: A Real Robot Benchmark for Streaming Task Understanding

model-releasesarxiv-cs-ro

arXiv:2609.36605v1 Announce Type: new Abstract: Understanding ongoing robot manipulation requires models to interpret visual observations in relation to interaction history and task progress. We intro

Tutorials

Scaling Influence Functions in LLMs through Eigenbasis-Corrected One-Bit Gradient Projection

arXiv:2609.37842v1 Announce Type: cross Abstract: Influence functions estimate how individual training examples affect the behavior of large language models (LLMs). Analyzing how training data influen

tutorialsarxiv-cs-ai

Industry

Sources: the US FTC is expanding a sweeping probe of Anthropic, OpenAI, and other frontier AI labs, and plans to issue formal demands to turn over information (New York Post)

New York Post: Sources: the US FTC is expanding a sweeping probe of Anthropic, OpenAI, and other frontier AI labs, and plans to issue formal demands to turn over information — WASHINGTON — The Federal

industrytechmeme

Think, Then Look: Active Spatial Reasoning for House-Scale 3D Scene Understanding

arXiv:2512.03284v2 Announce Type: replace Abstract: Spatial reasoning in large-scale 3D environments remains challenging for current vision--language models, which are typically constrained to room-sc

Thinking Before Thinking: Scaling Agentic Inference Through Meta-Reasoning

arXiv:2609.38147v1 Announce Type: new Abstract: As agents take on longer and more complex problems, controlling the execution becomes a task in its own right. Each step in the run brings new control c

UserProxyBench: Evaluating LLM User Simulators for Agent Benchmarks and Training

arXiv:2609.38043v1 Announce Type: new Abstract: Interactive agent benchmarks and multi-turn reinforcement learning increasingly place a second language model in the role of the user. This simulated us

Video2STL: Grounding VLM-Generated Temporal Specifications for Robot Learning

arXiv:2609.37519v1 Announce Type: cross Abstract: Video-based policy learning is particularly promising, as it illustrates target behaviors without requiring action annotations or embodiment-matched d

When Can Prefixes Compile LoRA? Exact Resource-Capped Tests for Frozen Attention

arXiv:2609.36766v1 Announce Type: new Abstract: Can a fixed continuous prefix replace a given low-rank adapter while the attention head stays frozen? In this research, we show that the answer depends

Agents

When Tools Silently Lie: Evaluating and Mitigating Blind Compliance in Tool-Augmented Data Agents

arXiv:2609.37153v1 Announce Type: new Abstract: Tool-augmented data agents rely on tool outputs for analytical decisions. Yet successful execution can return plausible but incorrect evidence, requirin

agentsarxiv-cs-ai

Who Warmed the Archives? LLMs Overestimate Historical Warmth

model-releasesarxiv-cs-cl

arXiv:2609.37499v1 Announce Type: new Abstract: Historical archives are an under-used source for extending the instrumental climate record backward in time, and LLMs offer a way to extract the indices

Active Causal Discovery Benchmark: Evaluating LLM Agents Under Budgeted Interventions

arXiv:2609.31675v1 Announce Type: new Abstract: We introduce the Active Causal Discovery Benchmark (ACDB), an SCM-grounded environment for evaluating whether LLM agents recover causal graph structure

Adapting Vision-Language Models for Human-Readable XAI in Industrial Object Detection

arXiv:2609.31690v1 Announce Type: new Abstract: Explainable Artificial Intelligence (XAI) solutions are essential for building trust in AI technologies and their integration in real manufacturing line

Safety

Attribution Without a Second Pass: Inline Per-Sample Gradient Provenance at ~1% Overhead

arXiv:2609.32380v1 Announce Type: new Abstract: Data attribution methods used in practice (TRAK, LoGRA, EK-FAC) are post-hoc: after training they make a second pass over the training set to recompute

safetyarxiv-cs-lg

Automated Species Identification in Camera Trap Images for Wildlife Conservation

arXiv:2609.35420v1 Announce Type: new Abstract: Wildlife conservation involves protecting, preserving, and managing wildlife species and their habitats. With today's rapid pace of human development, c

AutoRef: Harness Optimization for Agentic Multi-Reference Image Generation

arXiv:2609.35530v1 Announce Type: new Abstract: Recent image generation models can take multiple reference images as input and combine them into a new image. However, multi-reference image generation

CodeActionBench: Evaluating Agentic Code-as-Policy for Embodied Manipulation

model-releasesarxiv-cs-ro

arXiv:2609.33807v1 Announce Type: new Abstract: How well can general-purpose multimodal models turn visual understanding and reasoning into embodied manipulation via executable code? We introduce Code

Collaborative Principle Evolution via Evidence Transfer for Scientific Discovery

arXiv:2609.35315v1 Announce Type: new Abstract: Large Language Model (LLM)-based agents promise to automate scientific discovery, yet exploring the vast hypothesis space remains costly. Existing princ

Cost-free Spectral Estimation for Adaptive Newton--Schulz in Matrix Optimizers

arXiv:2609.33047v1 Announce Type: new Abstract: Matrix optimizers such as Muon transform each momentum matrix through an approximate orthogonalization, typically implemented by a small number of Newto

Depth Laws for the Precision Floor of Trained Neural Networks: Amplification, Residual Scaling, and a Quantization-Aware Training Paradox

arXiv:2609.32060v1 Announce Type: new Abstract: How many bits does a network need before its accuracy collapses, and how does this grow with depth? We study the precision floor, the perturbation level

From Attention Sensitivity to Layer Role: Revisiting Mixed-Precision Quantization of Transformers

arXiv:2609.34866v1 Announce Type: new Abstract: Most post-training quantization pipelines fit each weight matrix to its pretrained counterpart, one matrix at a time. Whether that proxy tracks what an

InfoEdit: Probing Global Layout Reasoning in Infographic Editing

arXiv:2609.33286v1 Announce Type: new Abstract: Multimodal foundation models edit natural photographs at production quality, yet the same models struggle with structured visual content such as infogra

Tutorials

Language Models for Text Classification: From Bag-of-Words to Jev

tutorialssebastian-raschka

The article discusses the newly released Jev AI model, which functions as a general‑purpose text classifier that can perform classification tasks more quickly and cheaply than large language models li

LLMs as Adaptive Meta-Solvers: Strategy-Diverse RL for Industrial-Scale Optimization

arXiv:2609.34427v1 Announce Type: new Abstract: Scaling LLM-based optimization from textbook-scale instances to real-world, industrial tasks remains a critical open challenge. Existing approaches are

Maintaining Benchmarks Against Increasingly Capable Agents: Detection and Remediation of Unearned Passes

arXiv:2609.34262v1 Announce Type: cross Abstract: Agentic benchmarks guide model selection and training. Yet an agent can pass a task without demonstrating the intended capability. Such outcomes const

MaLiang-Harness: A Programmable Path to Image and Video Generation

arXiv:2609.34309v1 Announce Type: new Abstract: Executable programs offer explicit control over how images and videos are constructed, but generating runnable code is only the beginning of visual crea

researcharxiv-cs-cv

Measure Learning at Steady State: A BIRD-SQL Formula 1 Case Study

arXiv:2609.31640v1 Announce Type: new Abstract: Continual Learning Bench scores learning as short-horizon gain versus a reset baseline and finds naive full-context ICL strongest among the memories it

MechBench: Can AI Scientific Agents Discover Mechanisms Beyond Phenomenal Laws?

arXiv:2609.35515v1 Announce Type: cross Abstract: Scientific discovery requires not only recovering mathematical laws that describe observable behavior, but also identifying the mechanisms that genera

Agents

NavHarness: Towards Lifelong Embodied Navigation

arXiv:2609.34276v1 Announce Type: new Abstract: Frontier models can now perform well on individual embodied navigation tasks through multi-round multimodal reasoning with simple tools. Across successi

agentsarxiv-cs-ro

OMP-MoE: Efficient Expert Pruning for Mixture-of-Experts LLMs via Orthogonal Matching Pursuit

arXiv:2609.31631v1 Announce Type: new Abstract: Mixture-of-Experts (MoE) models enable efficient scaling of large language models but face critical deployment challenges due to massive memory requirem

One Attack to Fool Them All: Highly Transferable Black-Box Adversarial Attacks on Frontier MLLMs

arXiv:2609.33833v1 Announce Type: new Abstract: Adversarial attacks have long posed a fundamental threat to machine learning systems. As multimodal large language models (MLLMs) rapidly evolve and bec

OpenAI launches Dots, its Muse competitor

model-releasesthe-verge-ai

OpenAI is responding to Meta's buzzy Muse AI with agentic helpers of its own: Dots. During its DevDay keynote on Tuesday, OpenAI announced that Dots will serve as always-on AI assistants that can 'do

Agents

Pre-registered tests of solid-state-physics-inspired LLM compression: a cluster-level negative result at small-language-model scale

arXiv:2609.34292v1 Announce Type: cross Abstract: We report a three-month autonomous research-agent program testing five solid-state-physics-inspired compression mappings on pretrained language models

agentsarxiv-cs-lg

Quantifying Behavioral Tails in Black-Box Language Models

arXiv:2609.33638v1 Announce Type: new Abstract: We introduce RareTrap, a framework for estimating the probability of severe behaviors in black box large language models (LLMs). A key challenge for pro

Query Expansion and Key Specialization in Transformer Attention Geometry

arXiv:2609.34273v1 Announce Type: cross Abstract: The projection of queries and keys are central to the attention mechanism in Transformer architectures. While they are mathematically symmetric, they

RAISE: Reinforcing Access Control Policy Synthesis in LLMs via Symbolic Evaluation

arXiv:2609.33796v1 Announce Type: cross Abstract: Translating natural-language access-control requirements into policies requires careful reasoning about permissions, constraints, and exceptions, and

Reflection 70B was released two years ago (September 2024)

model-releasesr-localllama

You may think that jev, OpenClaw or TurboQuant are super cool, but actually the coolest LLM invention happened two years ago As we all know, the best source of reliable information about LLMs is YouTu

Rethinking Circuit Evaluation: Do Circuits Explain Model Errors?

arXiv:2609.35686v1 Announce Type: new Abstract: Mechanistic interpretability (MI) aims to explain a model's behaviour through analyzing its internal computations; circuit-based explanations aim to iso

Safety

RubiCap: Rubric-Guided Reinforcement Learning for Dense Image Captioning

arXiv:2603.09160v2 Announce Type: replace Abstract: Dense image captioning is critical for vision-language pretraining and for text-to-image generation, but scaling expert-quality annotations is prohi

safetyarxiv-cs-cv

Share-Borne AI Virus: Memory-Hopping Attacks Across LLM Agents

arXiv:2609.35576v1 Announce Type: cross Abstract: Large language models are increasingly deployed as stateful assistants that retain information across interactions and use tools to read, modify, and

SolveEdit: Benchmarking Visual Problem Solving in Generative Models

arXiv:2609.35504v1 Announce Type: new Abstract: Machine intelligence is often evaluated through abstract reasoning problems, yet many real-world problems are visual, such as arranging objects, repairi

SRHarness: A Harness for Agentic Symbolic Regression

arXiv:2609.35501v1 Announce Type: cross Abstract: Recent agentic symbolic regression approaches increasingly rely on large language models to analyze data, select scientific operations, and refine hyp

Steering Language Model Goals with Value Transplant

arXiv:2609.34056v1 Announce Type: new Abstract: Reasoning models often act as if they pursue goals, but their efforts are not always directed toward what users intend, sometimes leading them to pursue

Timeline-Bench: Evaluating Agents on Realistic Video-Editing Tasks, from Raw Footage to Final Cut

arXiv:2609.35143v1 Announce Type: new Abstract: AI agents increasingly carry out long-horizon professional work, but their evaluations rarely require a finished creative deliverable. To this end, we i