>>213790579
there's usually fixed versions of LLMs. for example deepseek-r1-1776 removes the communism censorship from deepseek

whenever a new model is released you can usually find a fixed version in huggingface

>>213580149
i saw some vids of people trying openai's new 20b model and everyone's saying it's shit

also chatgpt fails pic related, openai is fumbling it hard. theyre irrelevant in LLMs pretty much

>>213532517
huggingface is french?
very good site i had no idea.

that being said europe gets mogged by china in LLMs very hard. no euro llm can perform basic tasks every chinese one can