Add benchmark methodology disclosure

Remove deprecated legacy GGUF: picolm-v2.1-81m-instruct-q4_k_m.gguf

Remove deprecated legacy GGUF: picolm-v2.1-81m-instruct-fp16.gguf

Fix: Update GGUF with valid llama-bpe pre-tokenizer metadata (picolm-v2.1-81m-instruct-q4_k_m.gguf)

Fix: Update GGUF with valid llama-bpe pre-tokenizer metadata (picolm-v2.1-81m-instruct-fp16.gguf)

Drop redundant untied lm_head.weight (Option 1)

Upload folder using huggingface_hub

initial commit