Search: “models”

If this prompt feels well written to you, it's because Suzanne is a writer in her little spare time! You can read her short story, Mall of A…

If this prompt feels well written to you, it's because Suzanne is a writer in her little spare time! You can read her short story, Mall of America here: https://suzannewang.com/mall-of-america It's on

model-releasesthariq--x

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

Model Releases

In early May, the best superforecasters predicted that, by the end of the year, the longest METR 80% task horizons would reach 3-4 hours. In…

DGX agent

In early May, the best superforecasters predicted that, by the end of the year, the longest METR 80% task horizons would reach 3-4 hours. In late May, Claude Mythos achieved that number. We also asked

model-releasesethan-mollick--x

3 Jun 2026

Jailbreak Attack Initializations as Extractors of Compliance Directions

arXiv:2502.09755v4 Announce Type: replace-cross Abstract: Safety-aligned LLMs respond to prompts with either compliance or refusal, each corresponding to distinct directions in the model's activation

safetyarxiv-cs-lg

Jerry Liu built one of the most installed pieces of AI plumbing of the last three years. Then he sat down and told me the framework era he h…

model-releasesjerry-liu--x

Jerry Liu built one of the most installed pieces of AI plumbing of the last three years. Then he sat down and told me the framework era he helped create is over. The agent harness ate the abstraction

Learning to Refine: Spectral-Decoupled Iterative Refinement Framework for Precipitation Nowcasting

arXiv:2606.02661v1 Announce Type: cross Abstract: Accurate precipitation nowcasting is vital for disaster mitigation, but deep learning methods face a key trade-off: regression models produce over-smo

researcharxiv-cs-ai

Learning to Solve, Forgetting to Retain: Correct-Set Turnover in RLVR

arXiv:2606.03087v1 Announce Type: new Abstract: Reinforcement learning with verifiable rewards (RLVR) improves the ability of large language model, yet headline accuracy gains often conceal a hidden c

Tutorials

Lethe: Adapter-Augmented Dual-Stream Update for Persistent Knowledge Erasure in Federated Unlearning

arXiv:2601.22601v2 Announce Type: replace Abstract: Federated unlearning (FU) aims to erase designated client-level, class-level, or sample-level knowledge from a global model. Existing studies common

tutorialsarxiv-cs-lg

Letting Tutor Personas Speak Up for LLMs: Learning Steering Vectors from Dialogue via Preference Optimization

arXiv:2602.07639v2 Announce Type: replace Abstract: With the emergence of large language models (LLMs) as a powerful class of generative artificial intelligence (AI), their use in tutoring has become

safetyarxiv-cs-cl

Locality Does Not Imply Reachability: Boundary Repair in Block-Sparse Causal Attention

arXiv:2606.02680v1 Announce Type: new Abstract: Sparse causal attention is usually described by sequence locality: nearby tokens should remain easy to access, while distant tokens may be dropped to re

Tools

M3 brings sparse attention + 1M context + multimodality, and Together did the hard serving work to make it fast. Great collaboration with th…

Anthropic's Claude 3.5 Sonnet (M3) model features sparse attention mechanisms, 1 million token context length, and multimodal capabilities, with Together AI optimizing the serving infrastructure to en

toolstogether-ai--x

Making Brain-Computer Interfaces More Secure

arXiv:2606.02597v1 Announce Type: new Abstract: The development of brain-computer interfaces (BCIs) based on electroencephalograms (EEGs) has advanced significantly mainly to machine learning. Althoug

Agents

MariData: One-Step Unpaired Image Translation for Maritime Environments

arXiv:2606.03246v1 Announce Type: new Abstract: The development on robust perception systems for Maritime Autonomous Surface Ships (MASS) is heavily constrained by the scarcity of diverse training dat

agentsarxiv-cs-cv

MARIO: Motion-Augmented Real-Time Multi-Sensor Inertial Odometry

model-releasesarxiv-cs-cv

arXiv:2606.02996v1 Announce Type: cross Abstract: Inertial odometry (IO) using only Inertial Measurement Units (IMUs) provides a lightweight solution for human motion tracking in augmented reality (AR

Mitigating False Credit Propagation: Probabilistic Graphical Reward Aggregation for Rubric-Based Reinforcement Learning

arXiv:2606.03361v1 Announce Type: new Abstract: Rubric-based rewards are increasingly used for open-ended language model post-training, but criterion-level scores are often aggregated as independent u

safetyarxiv-cs-lg

Mixed-Modality Dual Face-Hair Retrieval

model-releasesarxiv-cs-cv

arXiv:2606.03470v1 Announce Type: new Abstract: We introduce Dual Face-Hair Retrieval (DFHR), a new mixed-modality dual-reference task in image retrieval where a query consists of a face image specify

MultiTurnPSB: Evaluating Multi-Turn Jailbreak Attacks an dClassifier-Based Defenses for Medical AI Safety

arXiv:2606.02630v1 Announce Type: cross Abstract: Patient-facing medical chatbots are commonly evaluated on single-turn prompts, yet real users push back after refusals, add urgency, and invoke author

My timeline seems to have people surprised that U Chicago is getting Claude, but tons of schools (including U Penn where I teach) have schoo…

model-releasesethan-mollick--x

My timeline seems to have people surprised that U Chicago is getting Claude, but tons of schools (including U Penn where I teach) have school-wide AI There are lots of things that need to be figured o

OASIS: Outlier-Aware LUT-Based GEMM with Dual-Side Quantization for LLM Inference Acceleration

arXiv:2507.23035v4 Announce Type: replace Abstract: Large language models (LLMs) have demonstrated impressive capabilities across a wide range of applications, but demand substantial memory and comput

OpenAI ran a hiring challenge, but the top candidate was one they couldn’t hire: our autonomous research agent, Aiden. In Parameter Golf, Ai…

model-releasesclem-delangue--x

OpenAI ran a hiring challenge, but the top candidate was one they couldn’t hire: our autonomous research agent, Aiden. In Parameter Golf, Aiden ran for 22 days, and out-outperformed all 1,016 other re

Optimal Initialization in Depth: Lyapunov Initialization and Limit Theorems for Deep Leaky ReLU Networks

arXiv:2602.10949v2 Announce Type: replace-cross Abstract: Effective initialization in deep networks requires an understanding of random neural networks. In this work, a rigorous probabilistic analysis

Optimizing Explicit Unit-Distance Lower-Bound Certificates

arXiv:2606.03419v1 Announce Type: cross Abstract: The 2026 disproof of Erdos's unit-distance conjecture and Sawin's subsequent explicit quantitative refinement show that the maximum number u(n) of uni

Pathway-Structured Privileged Distillation for Deployable Computational Pathology

arXiv:2606.02877v1 Announce Type: new Abstract: Integrating transcriptomics and histopathology can improve cancer risk modelling, yet practical use is constrained by the limited availability of RNA pr

safetyarxiv-cs-cv

PerchRL: Vision-Based Agile Perching on Inclined Platforms under Rapid and Irregular Motion

arXiv:2606.03441v1 Announce Type: cross Abstract: Autonomous vision-based perching of quadrotors on moving inclined platforms is critical for air-ground collaboration but remains challenging due to th

PINNfluence: Interpreting PINNs through Influence Functions

arXiv:2409.08958v3 Announce Type: replace-cross Abstract: Physics-informed neural networks (PINNs) have emerged as a powerful deep learning approach for solving partial differential equations (PDEs) i

Plan2Map: A Multimodal Benchmark for Document-Grounded Geospatial Boundary Reconstruction from Planning Records

arXiv:2606.02747v1 Announce Type: cross Abstract: Planning records define restrictions over geographic areas, but their source documents often provide only indirect spatial evidence rather than machin

Planning with Uncertainty: Symmetries, Policy Inference, and Solution Compression

arXiv:2403.19883v2 Announce Type: replace Abstract: Fully-observable non-deterministic (FOND) planning is at the core of artificial intelligence planning with uncertainty. It models uncertainty throug

safetyarxiv-cs-ai

Position: Adversarial ML for LLMs Is Not Making Any Progress

arXiv:2502.02260v2 Announce Type: replace Abstract: In the past decade, considerable research effort has been devoted to securing machine learning (ML) models that operate in adversarial settings. Yet

Industry

Powering the Inference Era: Inside the DigitalOcean Data & Learning Layer

DigitalOcean discusses infrastructure and platform capabilities designed to support the inference phase of machine learning, where trained models are deployed to make predictions on new data at scale.

industrydigitalocean

ProtocolBench: Which LLM MultiAgent Protocol to Choose?

arXiv:2510.17149v3 Announce Type: replace Abstract: As large-scale multi-agent systems evolve, the communication protocol layer has become a critical yet under-evaluated factor shaping performance and

PsychoPass: Geometric Profiling of Multi-Turn Adversarial LLM Conversations

arXiv:2606.03136v1 Announce Type: cross Abstract: Multi-turn jailbreak attacks on large language models (LLMs) reveal a mismatch in current guardrails: they operate on individual turns, while attacks

safetyarxiv-cs-cl

Agents

Reproducibility is the New Copyleft: Defining AGI-oriented Reproducible Builds

arXiv:2606.03019v1 Announce Type: cross Abstract: Copyleft, as implemented in licenses such as the GNU General Public License, was a legal hack that used copyright to guarantee user freedom by tying t

agentsarxiv-cs-ai

Rethinking the Idiomaticity Decomposability Hypothesis: Evidence from Distributional Learning

arXiv:2606.03817v1 Announce Type: new Abstract: Idioms can be analysed in terms of their decomposability, the extent to which constituent meanings contribute to the figurative whole. Decomposability i

researcharxiv-cs-cl

Right Makes Might: Aligning Verified Hidden States Empowers RL Reasoning

arXiv:2606.03234v1 Announce Type: new Abstract: Reinforcement Learning from Verifiable Rewards (RLVR) has become the dominant approach for improving mathematical reasoning in large language models, ye

safetyarxiv-cs-lg

ROBUST-WT: Robust Uncertainty-aware Segmentation Transform via Whitening and Training Enhancements

arXiv:2606.03069v1 Announce Type: cross Abstract: Generalized segmentation of medical images prevents performance degradation when different imaging devices and clinical protocols are used across mult

Scalable Uncertainty Quantification for Extreme Weather Forecasting via Empirical Neural Tangent Kernels

arXiv:2606.02886v1 Announce Type: cross Abstract: Deep learning weather models now match numerical weather prediction accuracy while running orders of magnitude faster, but produce deterministic forec

researcharxiv-cs-ai

ScoreStop: Gradient-based early stopping using functional score tests

arXiv:2606.02740v1 Announce Type: cross Abstract: Gradient boosted decision trees require a stopping rule to avoid overfitting. The standard rule monitors a validation loss and stops if the loss fails

shower thought If: 1. AI is smarter than humans at law, therapy, etc. 2. Humans still like talking to other humans. Then: Humans are just an…

model-releasesjerry-liu--x

shower thought If: 1. AI is smarter than humans at law, therapy, etc. 2. Humans still like talking to other humans. Then: Humans are just an AI wrapper. Everyone should just regurgitate what Claude te

Sources: Benchmark raised 2B across two new funds, including a 1.25B fund for late-stage bets, its first growth fund after decades focusing on new startups (Kate Clark/Wall Street Journal)

Kate Clark / Wall Street Journal: Sources: Benchmark raised 2B across two new funds, including a 1.25B fund for late-stage bets, its first growth fund after decades focusing on new startups — After a

model-releasestechmeme

Sources: DeepSeek is set to raise ~7.4B in its first funding round from investors including Tencent and CATL at a valuation of between ~52B and ~$59B (Reuters)

Reuters: Sources: DeepSeek is set to raise ~7.4B in its first funding round from investors including Tencent and CATL at a valuation of between ~52B and ~59B — Chinese AI startup DeepSeek is set to ra

model-releasestechmeme

Local Ai

Tailoring Strictly Proper Scoring Rules for Downstream Tasks: An Application to Causal Inference

arXiv:2606.03332v1 Announce Type: new Abstract: Probabilistic models are typically trained using task-agnostic objectives like log-loss, which can lead to significant errors in downstream estimation.

local-aiarxiv-cs-lg

Temporal Action Selection for Action Chunking

arXiv:2511.04421v2 Announce Type: replace Abstract: Action chunking is a widely adopted approach in Learning from Demonstration (LfD). By modeling multi-step action chunks rather than single-step acti

safetyarxiv-cs-ro

TeX-1500: A Paired Real-World LWIR Hyperspectral Dataset and Benchmark for Temperature-Emissivity-Texture Decomposition

model-releasesarxiv-cs-cv

arXiv:2606.03806v1 Announce Type: new Abstract: Temperature-emissivity-texture (TeX) decomposition seeks to recover object heat state, material spectral response, and visible-like geometric texture fr

Text-attributed Graph Condensation via Text Selection and Attribute Matching

arXiv:2606.03839v1 Announce Type: new Abstract: Text-Attributed Graph (TAG) is an important type of graph structured data, where each node has a text description. TAG models usually train a Graph Neur

The Geometry of LLM-as-Judge: Why Inter-LLM Consensus Is Not Human Alignment

model-releasesarxiv-cs-cl

arXiv:2606.03043v1 Announce Type: new Abstract: LMs-as-judges are now standard, yet judges agree strongly with one another while agreeing only weakly with humans. We test whether this reflects shared

The Impact of Configuring Agentic AI Coding Tools on Build-vs-Buy Decisions: A Study Protocol

arXiv:2606.03907v1 Announce Type: cross Abstract: Agentic AI coding tools write code with increasing autonomy and in doing so decide when to import a library and when to implement functionality from s

The Ringelmann Effect in Multi-Agent LLM Systems: A Scaling Law for Effective Team Size

arXiv:2606.02646v1 Announce Type: cross Abstract: Inference-time multi-agent LLM scaling lacks a shared unit: counting nominal agents conflates cost with independent evidence. We derive a two-paramete

The Shadow Price of Reasoning: Economic Perspective on Optimal Budget Allocation for LLMs

arXiv:2606.03092v1 Announce Type: new Abstract: Inference-time scaling has emerged as a critical avenue for enhancing Large Language Models' performance, yet real-world deployment is constrained by st

safetyarxiv-cs-ai

This is really intellectually dishonest. I have not been arguing that LLM token prices are increasing (though the all you can eat buffet is …

model-releasesgary-marcus--x

This is really intellectually dishonest. I have not been arguing that LLM token prices are increasing (though the all you can eat buffet is over), I have been arguing the *opposite*, viz that they wil