onebench / README.md
librarian-bot's picture
Librarian Bot: Add language metadata for dataset
f6508df verified
|
raw
history blame
561 Bytes
metadata
language:
  - en
dataset_info:
  features:
    - name: text
      dtype: string
  splits:
    - name: train
      num_bytes: 845359
      num_examples: 7522
  download_size: 640394
  dataset_size: 845359
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*

https://github.com/aicrumb/one-bench

lower = better

Pythia 410m: 881.89
Pythia 1B: 192.55
Pythia 1.4B: 220.02
Pythia 2.8B: 471.30
Pythia 6.7B: 783.43
Mistral v0.1 7B: 862.12
RWKV-4 430m: 327.53
RWKV-4 1b5: 273.75
OPT-1.3B: 143.02
OPT-2.7B: 60.97
OPT-6.7B: 111.03