Fairy2i: Training Complex LLMs from Real LLMs with All Parameters in {pm 1, pm i}
Open paper
Benchmark
Model
Rank
Results
common-sense-reasoning-on-arc-easy
Fairy2i-W2 (2-bit)
#48
Accuracy: 72.73
question-answering-on-piqa
Fairy2i-W2 (2-bit)
#105
Accuracy: 76.17
sentence-completion-on-hellaswag
Fairy2i-W2 (2-bit)
#171
Accuracy: 53.33
Rank counts only results with a code link.