open-llm

A tweet announcing a torrent-based site for downloading open LLM model weights, positioning it as a pirate bay alternative for open models.

The poster states that the MoQ GGUFs of the LFM2.5 8B A1B model offer the best accuracy-to-size ratio, advising against using versions with less than 95% accuracy recovery.

This paper measures maximum activation magnitudes across 27 checkpoints from 8 open LLM families, finding significant variance across families, architectures, and training stages, with implications for low-bit quantization and deployment.