strategy-scope/si_et_al-ideation-gpt5mini-full-20260415_012526-evaluation
si_et_al-ideation-gpt5mini-full-20260415_012526-evaluation Aggregate per-topic metrics (+ overall row where topic = __overall__). Parent: si_et_al-ideation-gpt5mini-full-20260415_012526 Run Parameters model: gpt-5-mini judge_model: anthropic/claude-sonnet-4-5-20250929 strategy: full n_ideas_per_run: 5 n_runs: 3 n_topics: 7 rag: False Columns topic / n_ideas mean_overall: mean of per-idea 1-10 overall scores (port of… See the full description on the dataset page: https://huggingface.co/datasets/strategy-scope/si_et_al-ideation-gpt5mini-full-20260415_012526-evaluation.
03
sietal-ideation-gpt5mini-full-20260415_012526-evaluation
Aggregate per-topic metrics (+ overall row where topic = __overall__). Parent: si_et_al-ideation-gpt5mini-full-20260415_012526
Run Parameters
- model: gpt-5-mini
- judge_model: anthropic/claude-sonnet-4-5-20250929
- strategy: full
- n_ideas_per_run: 5
- n_runs: 3
- n_topics: 7
- rag: False
Columns
- topic / n_ideas
- mean_overall: mean of per-idea 1-10 overall scores (port of airesearcher/src/ideadirect_score.py)
- diversity_cosine_{pairwise,nn} / diversity_jaccard_{pairwise,nn}: 1 - similarity
