LLM 5
- Why Modern LLMs Dropped Mean Centering (And Got Away With It)
- The Epsilon Trap: When Adam Stops Being Adam
- Entropic Instruction Following: Does Semantic Coherence Help LLMs Follow Instructions?
- Elements Of Mechanistic Interpretability: From Observation to Causation
- SFT vs. DPO (/ RLHF)- A Visual Guide to What Your LLM Actually Learns