Unboxing LLMs > loading...

Author: Rakshit Kalra

Home / Archives

Latest Posts

November 27, 2024

Working For a Self-Saboteur – The Uncooperative Objective Function

November 5, 2024

Stealing User Prompts from Mixture‑of‑Experts: How Tiebreaks Become a Side‑Channel

November 2, 2024

Uncovering Mesa-Optimization in Transformers

October 31, 2024

Go Zero‑Shot: Faster & More Reproducible LLM Benchmarking

October 28, 2024

STRING: A Pragmatic Fix for LLMs’ Long‑Context Blind‑Spots

October 14, 2024

Buffer of Thoughts – Thought‑Augmented Reasoning for LLMs