List your company

Creative writing on Mythos 100 prompts (test)

834Word CountGPT-5
281.76425.13568.5711.87May 23, 2025

Evaluation results

MethodLinks
GPT-5Generation setting=Vanilla Baseline2025.05834—1001001.711.030.194.20
Qwen-3-32BGeneration setting=Vanilla Baseline2025.05699—10010015.860.183.220
GPT-5Generation setting=Frankentext, Snippets=1.5k2025.056758292422.764.340.215.8879
Gemini 2.5 ProGeneration setting=Vanilla Baseline2025.05593—1001001.766.410.193.180
Qwen-3-32BGeneration setting=Frankentext, Snippets=1.5k2025.055783691542.21.370.184.027
Deepseek-R1Generation setting=Vanilla Baseline2025.05550—1001001.283.490.24.130
Gemini-2.5-ProGeneration setting=RAG Baseline, Snippets=1.5k2025.055380.63100991.566.430.23.462
Gemini 2.5 ProGeneration setting=Frankentext, Snippets=1.5k2025.0552175100812.749.270.224.2159
Claude-4-SonnetGeneration setting=Vanilla Baseline2025.05477—1001001.41.70.183.310
Gemini 2.5 ProGeneration setting=Frankentext (Ablation), Snippets=5k2025.054517997852.789.480.215.1372
Gemini 2.5 ProGeneration setting=Frankentext (Ablation), Snippets=10k2025.054487899852.819.120.215.4370
Claude-4-SonnetGeneration setting=Frankentext, Snippets=1.5k2025.053175198862.650.193.9947
Deepseek-R1Generation setting=Frankentext, Snippets=1.5k2025.053034291722.798.310.24.6623