Experiment Result Viewer (50 examples)
Instruction and reference are given to an evaluator machine.
The evaluator machine selected the better answer by pairwise comparisons.
You can select model outputs with the Model Selector and download them in JSON format.
🎯 Evaluator
alpaca
baize
chatgpt
cohere
dolly
falcon
gpt4
instructgpt
koala
llama
llama v2
mpt
open assistant
red pajama
vicuna
wizard lm
human
🎨 Model Selector
all
alpaca
baize
chatgpt
cohere
dolly
falcon
gpt4
instructgpt
koala
llama
llamav2
mpt
open_assistant
red_pajama
vicuna
wizard_lm
🎲 Current Evaluator: alpaca
Instruction
Reference
Ranking
Instruction
Reference
Ranking
Instruction
Reference
Ranking
Instruction
Reference
Ranking
Instruction
Reference
Ranking
💾 Download Result