bert-base-nli-mean-tokens/sentence_bert_config.json

{
  "max_seq_length": 128,
  "do_lower_case": false
}