README.md

metadata

license: apache-2.0
task_categories:
  - question-answering
  - text-generation
  - sentence-similarity
language:
  - en
pretty_name: FLAN v2 Cot Deduplicated Dataset
size_categories:
  - 1K<n<10K

FLAN v2 Cot Deduplicated Dataset

Data Preprocessing

• Remove instructions with less than 100 tokens in 'targets'.
• Dedepulicate Dataset using cosine similarity with a threshold of 0.95.

Code

• Github repo : https://github.com/AJlearner46/Deduplicate-flanv2-finetune-LLaMa3-

Acknowledgments

• The original dataset is provided by SirNeural/flan_v2.
• Tokenizer used: bert-base-uncased from Hugging Face.