Hacker News headlines

30

SPy: An interpreter and compiler for a fast statically typed variant of Python (antocuni.eu)

a day ago | og_kalu | antocuni.eu | frontpage

11

Emergent Introspective Awareness in Large Language Models (transformer-circuits.pub)

a week ago | og_kalu | transformer-circuits.pub | best

1

Quantifying the algorithmic improvement from reasoning models (epoch.ai)

3 months ago | og_kalu | epoch.ai | newest

1

Evidence of interrelated cognitive-like capabilities in large language models (sciencedirect.com)

5 months ago | og_kalu | sciencedirect.com | newest

9

Atlas: Learning to Optimally Memorize the Context at Test Time (arxiv.org)

6 months ago | og_kalu | arxiv.org | frontpage

15

Gemini Diffusion (deepmind.google)

6 months ago | og_kalu | deepmind.google | frontpage

3

Tails Tell Tales: Chapter-Wide Manga Transcriptions with Character Names (arxiv.org)

9 months ago | og_kalu | arxiv.org | newest

2

Over-Tokenized Transformer: Vocabulary Is Generally Worth Scaling (arxiv.org)

9 months ago | og_kalu | arxiv.org | newest

2

LLMs struggle with perception, not reasoning, in ARC-AGI (anokas.substack.com)

9 months ago | og_kalu | substack.com | newest

2

EvaByte: Efficient Byte-Level Language Models at Scale (hkunlp.github.io)

10 months ago | og_kalu | github.io | newest

1

Tell me about yourself: LLMs are aware of their learned behaviors (arxiv.org)

10 months ago | og_kalu | arxiv.org | newest

2

Imagine While Reasoning in Space: Multimodal Visualization-of-Thought (arxiv.org)

10 months ago | og_kalu | arxiv.org | newest

1

LLMs struggle with perception, not reasoning, in ARC-AGI (anokas.substack.com)

10 months ago | og_kalu | substack.com | newest

2

Byte Latent Transformer: Patches Scale Better Than Tokens (meta.com)

11 months ago | og_kalu | meta.com | newest

1

Mastering Board Games by External and Internal Planning with Language Models (deepmind.google)

11 months ago | og_kalu | deepmind.google | newest

2

Emergence of Hidden Capabilities: Exploring Learning Dynamics in Concept Space (arxiv.org)

12 months ago | og_kalu | arxiv.org | newest

3

GameGen-X: Open-World Video Game Generation (gamegen-x.github.io)

a year ago | og_kalu | github.io | newest

45

TokenFormer: Rethinking Transformer Scaling with Tokenized Model Parameters (arxiv.org)

a year ago | og_kalu | arxiv.org | best

4

Kurzgesagt: We Fell for the Oldest Lie on the Internet [video] (youtube.com)

a year ago | og_kalu | youtube.com | newest

1

Relaxed Recursive Transformers: Effective Parameter Sharing with Layer-Wise LoRA (arxiv.org)

a year ago | og_kalu | arxiv.org | newest

2

Solving Global Lyapunov functions: open problem in mathematics with transformers (arxiv.org)

a year ago | og_kalu | arxiv.org | newest

2

ChatGPT Topped 3B Visits in September (similarweb.com)

a year ago | og_kalu | similarweb.com | newest

1

Tx-LLM: Supporting therapeutic development with large language models (research.google)

a year ago | og_kalu | research.google | newest

1

Tx-LLM: Supporting therapeutic development with large language models (research.google)

a year ago | og_kalu | research.google | newest

2

Visual Autoregressive Modeling: Image Generation via Next-Resolution Prediction (arxiv.org)

a year ago | og_kalu | arxiv.org | newest

4

xAI's Colossus (100k H100 cluster) has begun training (twitter.com/elonmusk)

a year ago | og_kalu | twitter.com | newest

1

Recite, Reconstruct, Recollect: Memorization in LMs as a Multifaceted Phenomenon (arxiv.org)

a year ago | og_kalu | arxiv.org | newest

1

GPT-4o's image generation capabilities (twitter.com/gdb)

a year ago | og_kalu | twitter.com | newest

1

LLMs for few-shot low level robot control by representing trajectories as tokens (twitter.com/ed__johns)

a year ago | og_kalu | twitter.com | newest

1

Keypoint Action Tokens Enable In-Context Imitation Learning in Robotics (robot-learning.uk)

a year ago | og_kalu | robot-learning.uk | newest

2

You can now edit DALL·E images in ChatGPT (twitter.com/openai)

a year ago | og_kalu | twitter.com | newest

6

Microsoft and Open AI Plot $100B AI Supercomputer Called "Stargate" (reuters.com)

a year ago | og_kalu | reuters.com | newest

1

Arrows of Time for Large Language Models (arxiv.org)

a year ago | og_kalu | arxiv.org | newest

1

3D Vision-Language-Action Generative World Model (umass.edu)

a year ago | og_kalu | umass.edu | newest

2

Introducng RFM-1: Giving robots human-like reasoning capabilities (covariant.ai)

a year ago | og_kalu | covariant.ai | newest

1

LLMs and the Abstraction and Reasoning Corpus (arxiv.org)

a year ago | og_kalu | arxiv.org | newest

2

Beyond Language Models: Byte Models Are Digital World Simulators (byte-gpt.github.io)

a year ago | og_kalu | github.io | newest

1

A Vision Check-Up for Language Models (csail.mit.edu)

a year ago | og_kalu | mit.edu | newest

1

The Impact of Reasoning Step Length on Large Language Models (arxiv.org)

a year ago | og_kalu | arxiv.org | newest

36

Chain-of-Thought Reasoning Without Prompting (arxiv.org)

a year ago | og_kalu | arxiv.org | best

2

The Manga Whisperer: Automatically Generating Transcriptions for Comics (github.com/ragavsachdeva)

a year ago | og_kalu | github.com | newest

2

The Manga Whisperer: Automatically Generating Transcriptions for Comics (github.com/ragavsachdeva)

a year ago | og_kalu | github.com | newest

1

Show HN: Automatic Translation of Comics (Bande Dessinée, Manga, Webtoons, etc.) (github.com/ogkalu2)

a year ago | og_kalu | github.com | newest

1

Towards General World Models for Video Generation via Predicting Masked Tokens (world-dreamer.github.io)

a year ago | og_kalu | github.io | newest

1

Towards Conversational Diagnostic AI (arxiv.org)

a year ago | og_kalu | arxiv.org | newest

43

OpenAI Announces $10M Superalignment Grants (openai.com)

a year ago | og_kalu | openai.com | newest

1

From Text to Motion: Grounding GPT-4 in a Humanoid Robot "Alter3" (tnoinkwms.github.io)

a year ago | og_kalu | github.io | newest

1

Using Large Language Models for Hyperparameter Optimization (arxiv.org)

a year ago | og_kalu | arxiv.org | newest

2

Scaling Transformers for skillful and reliable medium-range weather forecasting (arxiv.org)

a year ago | og_kalu | arxiv.org | newest

47

Sequential modeling enables scalable learning for large vision models (yutongbai.com)

a year ago | og_kalu | yutongbai.com | best

1

Subject/Style-driven image generation and Audio editing with a Multimodal LLM (codi-2.github.io)

a year ago | og_kalu | github.io | newest

2

CoDi-2: In-Context, Interleaved, and Interactive Any-to-Any Generation (codi-2.github.io)

a year ago | og_kalu | github.io | newest

2

Max Tegmark on Computation, Substrate-Independence and Consciousness (edge.org)

a year ago | og_kalu | edge.org | newest

4

LLMs and the Abstraction and Reasoning Corpus (arxiv.org)

a year ago | og_kalu | arxiv.org | newest

55

Misalignment and Deception by an autonomous stock trading LLM agent (arxiv.org)

a year ago | og_kalu | arxiv.org | best

5

LLMs and the Abstraction and Reasoning Corpus (arxiv.org)

a year ago | og_kalu | arxiv.org | newest

2

Unprompted, LLM Agents can strategically deceive users when put under pressure (arxiv.org)

a year ago | og_kalu | arxiv.org | newest

1

Brains, Planes, Blimps and Algorithms (reddit.com)

a year ago | og_kalu | reddit.com | newest

1

A conceptual precursor to today's language machines (hedgehogreview.com)

a year ago | og_kalu | hedgehogreview.com | newest

2

Large Language Models Can Strategically Deceive Their Users When Under Pressure (arxiv.org)

a year ago | og_kalu | arxiv.org | newest

7

Jarvis-1: Open-World Multi-Task Agents with Memory-Augmented Multimodal LLMs (craftjarvis-jarvis1.github.io)

a year ago | og_kalu | github.io | best

1

RoboVQA: Multimodal Long-Horizon Reasoning for Robotics (robovqa.github.io)

a year ago | og_kalu | github.io | newest

2

Set-of-Mark Prompting Unleashes Extraordinary Visual Grounding in GPT-4Vision (som-gpt4v.github.io)

a year ago | og_kalu | github.io | newest

1

The Dark Side of Antarctica [video] (youtube.com)

a year ago | og_kalu | youtube.com | newest

1

Taken out of context: On measuring situational awareness in LLMs (arxiv.org)

2 years ago | og_kalu | arxiv.org | newest

1

Interactive Robot Learning from Verbal Correction (ut-austin-rpl.github.io)

2 years ago | og_kalu | github.io | newest

1

Unleashing the Power of Pre-Trained LLMs for Offline Reinforcement Learning (arxiv.org)

2 years ago | og_kalu | arxiv.org | newest

4

CodeFusion: A Pre-Trained Diffusion Model for Code Generation (huggingface.co)

2 years ago | og_kalu | huggingface.co | newest

1

Learning in High Dimension Always Amounts to Extrapolation (arxiv.org)

2 years ago | og_kalu | arxiv.org | newest

2

Multi-Game Decision Transformers (sites.google.com)

2 years ago | og_kalu | google.com | newest

1

In-Context Learning Creates Task Vectors (arxiv.org)

2 years ago | og_kalu | arxiv.org | newest

8

LLMs playing chess are sensitive to how the position came to be (github.com/dpaleka)

2 years ago | og_kalu | github.com | frontpage

3

Comprehension of Sentences by Bottlenosed Dolphins (sciencedirect.com)

2 years ago | og_kalu | sciencedirect.com | newest

2

Is Computer Vision dead? (tenyks.ai)

2 years ago | og_kalu | tenyks.ai | newest

5

GPT-4 designs reward functions for robot dexterity at Super-human level (twitter.com/drjimfan)

2 years ago | og_kalu | twitter.com | frontpage

2

Eureka! Nvidia Research Breakthrough Puts New Spin on Robot Learning (nvidia.com)

2 years ago | og_kalu | nvidia.com | newest

2

Eureka: Human-Level Reward Design via Coding Large Language Models (arxiv.org)

2 years ago | og_kalu | arxiv.org | newest

1

Revealing the structure of language model capabilities (arxiv.org)

2 years ago | og_kalu | arxiv.org | newest

10

85% of the variance in LLM performance is explained by a single factor, g (reddit.com)

2 years ago | og_kalu | reddit.com | frontpage

0

Character AI's Group Chat Feature (reddit.com)

2 years ago | og_kalu | reddit.com | newest

2

85% of the variance in LLM performance is explained by a single factor, g (reddit.com)

2 years ago | og_kalu | reddit.com | newest

1

Unveiling the General Intelligence Factor in Language Models (arxiv.org)

2 years ago | og_kalu | arxiv.org | newest

20

xVal: A continuous number encoding for large language models (arxiv.org)

2 years ago | og_kalu | arxiv.org | best

1

Inductive reasoning in humans and large language models (sciencedirect.com)

2 years ago | og_kalu | sciencedirect.com | newest

2

Identifying depression and its determinants: ChatGPT vs. primary care physicians (bmj.com)

2 years ago | og_kalu | bmj.com | newest

1

Using Transformers for Multi-Agent Reinforcement Learning (arxiv.org)

2 years ago | og_kalu | arxiv.org | newest

1

Multi-Agent Reinforcement Learning Is a Sequence Modeling Problem (arxiv.org)

2 years ago | og_kalu | arxiv.org | newest

1

Large Language Models Can Learn Rules (arxiv.org)

2 years ago | og_kalu | arxiv.org | newest

2

The Geometry of Truth: Emergent Linear Structure in How LLMs Represent Truth (arxiv.org)

2 years ago | og_kalu | arxiv.org | newest

1

Just Ask for Calibration: Eliciting Calibrated Confidence Scores from LLMs (arxiv.org)

2 years ago | og_kalu | arxiv.org | newest

1

Can LLMs provide useful feedback on research papers? A broad empirical analysis (arxiv.org)

2 years ago | og_kalu | arxiv.org | newest

68

Training language models with pause tokens (arxiv.org)

2 years ago | og_kalu | arxiv.org | best

1

Boolformer: Symbolic Regression of Logic Functions with Transformers (arxiv.org)

2 years ago | og_kalu | arxiv.org | newest

3

The Dawn of Large Multimodal Models: Preliminary Explorations with GPT-4V(ision) (arxiv.org)

2 years ago | og_kalu | arxiv.org | newest

1

Catch an AI Liar: Lie Detection in Black-Box LLMs by Asking Unrelated Questions (arxiv.org)

2 years ago | og_kalu | arxiv.org | newest

1

Large Language Models as Superpositions of Cultural Perspectives (arxiv.org)

2 years ago | og_kalu | arxiv.org | newest

2

DALL-E 3 will finish rolling out to all Bing Image users by 8PM PST today (twitter.com/mparakhin)

2 years ago | og_kalu | twitter.com | newest

2

Large Language Models Understand and Can Be Enhanced by Emotional Stimuli (arxiv.org)

2 years ago | og_kalu | arxiv.org | newest

3

The Internal State of a Large Language Model Knows When Its Lying (arxiv.org)

2 years ago | og_kalu | arxiv.org | newest

2

Dalle-3 Results and Requests (reddit.com)

2 years ago | og_kalu | reddit.com | newest