List your company

Creative Writing on Alpaca SFT Short Stories

3.7Self-BLEU (Diversity)λ-PR
2.8528.57614.320.024Apr 30, 2026

Evaluation results

MethodLinks
λ-PRBackbone=Mistral-12B2026.043.72.5
λ-PRBackbone=LLAMA-3.1-8B2026.043.82.7
λ-PRBackbone=OLMo-2-13B2026.043.93
λ-PRBackbone=PHI-4-14B2026.0443.1
λ-PRBackbone=Pythia-12B2026.044.12.1
λ-PRBackbone=QWEN-3-8B2026.0453.4
BASEBackbone=Pythia-12B2026.049.32.9
TOFUBackbone=Pythia-12B2026.0410.83.6
TOFUBackbone=LLAMA-3.1-8B2026.0411.24.5
BASEBackbone=Mistral-12B2026.0411.43.9
BASEBackbone=LLAMA-3.1-8B2026.0411.53.5
GEMBackbone=LLAMA-3.1-8B2026.0411.54.5
TOFUBackbone=OLMo-2-13B2026.0411.94.6
BASEBackbone=OLMo-2-13B2026.0412.33.7
TOFUBackbone=PHI-4-14B2026.0412.44.6
TOFUBackbone=Mistral-12B2026.0412.74.6
TOFUBackbone=QWEN-3-8B2026.0412.74.5
GEMBackbone=Pythia-12B2026.04133.4
GEMBackbone=OLMo-2-13B2026.0413.24.6
GEMBackbone=PHI-4-14B2026.0413.44.7
BASEBackbone=PHI-4-14B2026.0413.83.4
GEMBackbone=QWEN-3-8B2026.0413.94.5
GEMBackbone=Mistral-12B2026.0414.34.6
FLBackbone=LLAMA-3.1-8B2026.0414.34.5
FLBackbone=Pythia-12B2026.0414.63.8
FLBackbone=OLMo-2-13B2026.0415.94.7
FLBackbone=PHI-4-14B2026.0416.24.7
FLBackbone=QWEN-3-8B2026.0416.44.6
FLBackbone=Mistral-12B2026.0416.84.7
CEBackbone=QWEN-3-8B2026.0422.34.6
CEBackbone=Mistral-12B2026.0422.54.8
CEBackbone=LLAMA-3.1-8B2026.0422.64.7
CEBackbone=Pythia-12B2026.0423.83.7
CEBackbone=OLMo-2-13B2026.0424.54.8
CEBackbone=PHI-4-14B2026.0424.84.8
BASEBackbone=QWEN-3-8B2026.0424.92.9