Travis King

travisking's activity

• All
• Models
• Datasets
• Spaces
• Buckets
• Papers
• Collections
• Community
• Posts
• Upvotes
• Likes
• Articles

upvoted 2 papers 14 days ago

The Functionalizer: Lossless Functional Decomposition for Subword Tokenization

Paper • 2609.15991 • Published 20 days ago • 17

Gricea: An Open Science Platform for Conversational AI Research

Paper • 2609.22039 • Published 20 days ago • 13

upvoted 3 papers 22 days ago

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

Paper • 2609.15504 • Published 24 days ago • 37

Online Draft Co-Training for Speculative Decoding in Large-Scale, Long-Context RL Post-Training

Paper • 2609.07108 • Published about 1 month ago • 36

An Open Recipe for IMO Gold: Training Nemotron for Olympiad Mathematics

Paper • 2609.10712 • Published 29 days ago • 45

upvoted 2 papers about 1 month ago

Puro-2B: Poor Lab's Qwen2-1.5B Trained on RTX 5090 within $5090

Paper • 2608.27370 • Published Aug 27 • 40

Sliding-window beats linear attention

Paper • 2608.28444 • Published Aug 28 • 19

upvoted 3 papers about 2 months ago

FM-Bench: A Benchmark for Long-Horizon Management with Competing Agents

Paper • 2608.18423 • Published Aug 19 • 21

LLMRouter: Unified Infrastructure for Developing, Evaluating, and Deploying LLM Routers

Paper • 2608.06867 • Published Aug 7 • 114

Beyond Final Scores: A Systematic Evaluation of Agents for Long-Horizon AI Research and Development

Paper • 2608.13417 • Published Aug 13 • 59

upvoted an article 2 months ago

view article

Article

TutorMoments: Do AI tutors know when to help and when to hold back?

• 

allenai

•

Aug 7

• 31

upvoted 3 papers 2 months ago

Are the Financial Reasoning from LLMs Credible? A Real World Test over Long-Horizon Statements

Paper • 2607.28661 • Published Jul 22 • 19

When Attention Goes Blind: Numerical Failure in ALiBi Positional Encodings

Paper • 2608.03994 • Published Aug 4 • 9

MerchantBench: Benchmarking LLM Agents for Long-Term Coherence in E-Commerce Operations

Paper • 2607.28956 • Published Jul 31 • 114

upvoted a collection 2 months ago

Instella-MoE ✨

Collection

Family of fully open 16B MoE LLM with 2.8B active params per token, trained on AMD Instinct™ MI300 & MI325 GPUs. https://arxiv.org/abs/2609.00791 • 6 items • Updated 19 days ago • 19

upvoted a paper 2 months ago

LLMs Get Lost in Evolving User Intent

Paper • 2607.20734 • Published Jul 22 • 24

upvoted an article 3 months ago

view article

Article

NVIDIA Nemotron 3 Embed Ranks #1 Overall on RTEB, Advancing Agentic Retrieval

• 

nvidia

•

Jul 16

• 60

upvoted a collection 3 months ago

Nemotron 3 Embed

Collection

Open embedding models for enterprise RAG, agentic retrieval, code search, and agent memory. • 3 items • Updated Aug 11 • 41

upvoted 2 articles 3 months ago

view article

Article

Distillation in 2026 (so far): which frontier models use it and how

• 

sergiopaniego

•

Jul 8

• 22

view article

Article

Which tokens does a hybrid model predict better?

• 

allenai

•

Jun 25

• 9