Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Unconditional Text Generation on OpenWebText (Qualitative Metrics)
Loading...
3.41
Clarity
BD3-LM + Gumbel Distillation
2.4012
2.6631
2.925
3.1869
Mar 23, 2026
Clarity
Grammaticality
Factuality
Style
Creativity
Updated 25d ago
Evaluation Results
Method
Method
Links
Clarity
Grammaticality
Factuality
Style
Creativity
BD3-LM + Gumbel Distillation
LLM Judge=Gemini-2.5-pro
2026.03
3.41
3.22
3.78
3.35
2.68
BD3-LM
LLM Judge=Gemini-2.5-pro
2026.03
2.89
2.95
3.21
3.34
2.75
MDLM + Gumbel Distillation
LLM Judge=Gemini-2.5-pro
2026.03
2.86
2.57
3.31
2.57
2.36
MDLM
LLM Judge=Gemini-2.5-pro
2026.03
2.44
2.22
2.7
2.32
2.22
Feedback
Search any
task
Search any
task