Large Language Models as Tool Makers

Stronger Together: on the Articulation of Ethical Charters, Legal Tools, and Technical Documentation in ML

Intriguing Properties of Quantization at Scale

Improving Mathematical Reasoning with Process Supervision

AutoML for neuromorphic computing and application-driven co-design: asynchronous, massively parallel optimization of spiking architectures

Improving Factuality and Reasoning in Language Models through Multiagent Debate

Learning to Generate Novel Scientific Directions with Contextualized Literature-based Discovery

Deep learning-guided discovery of an antibiotic targeting Acinetobacter baumannii

is there a prompt science after engineering and design?

Resolving code review comments with ML

LIMA: Less Is More for Alignment

IndoorSim-to-OutdoorReal: Learning to navigate outdoors without any outdoor experience

To Compress or Not to Compress -- Self-Supervised Learning and Information Theory: A Review

A Guide to ICLR 2023 — 10 Topics and 50 papers you shouldn't miss

Visual Blocks for ML: Accelerating machine learning prototyping with interactive tools

Hyena Hierarchy: Towards Larger Convolutional Language Models

Language Models can Solve Computer Tasks

TPU v4: An Optically Reconfigurable Supercomputer for Machine Learning with Hardware Support for Embeddings

Visualizing the Implicit Model Selection Tradeoff

DiffDock: Diffusion Steps, Twists, and Turns for Molecular Docking

Google updates Bard to better answer math and logic questions, coding coming soon

Reflexion: an autonomous agent with dynamic memory and self-reflection

Magnushammer: A Transformer-based Approach to Premise Selection

Exphormer: Sparse Transformers for Graphs

Generative AI: Language, Images and Code - A Conversation with CSAIL

Eliciting Latent Predictions from Transformers with the Tuned Lens

Prompt Programming for Large Language Models: Beyond the Few-Shot Paradigm

Toolformer: Language Models Can Teach Themselves to Use Tools

Solving a machine-learning mystery

Progress measures for grokking via mechanistic interpretability