Stories by fzliu

Benchmark for Evaluating Text Embeddings

MAIR: A Benchmark for Evaluating Instructed Retrieval

Wanting to Be Understood Explains the Meta-Problem of Consciousness

The Gumbel-Softmax Distribution

The Value of Chess Pieces

AbstentionBench: Reasoning LLMs Fail on Unanswerable Questions

CURE: A Dataset for Clinical Understanding and Retrieval Evaluation

Benchmark for Evaluating Text Embeddings

Extracting memorized pieces of books from open-weight language models

The FCC Builds a Firewall Around US-Bound Electronics

Towards Understanding Sycophancy in Language Models

Extended Thinking Tips

Embedding Benchmark for Retrieval

Torch Backends

Hideo Kojima's Boss Fight with Time

AI Risk Repository

RAG for Contract Analysis

Embedding Benchmark for Retrieval

tau²-bench

Video Search with Multimodal AI

1
2
3
4
5
6
7
8
9
10