gpt-neox-japanese-2.7b / tokenizer_config.json
SO0529
first commit
88fe7f0
{
"pad_token": "<|endoftext|>",
"unk_token": "<|endoftext|>",
"bos_token": "<|startoftext|>",
"eos_token": "<|endoftext|>",
"do_clean_text": false
}