iol-ai-2026-solver / experiments_log.md
Lucia Domenichelli
Improve answer parsing and validation
97fc3d1 verified
|
Raw
History Blame Contribute Delete
362 Bytes
# IOL-AI 2026 experiments
## baseline_accepted
- Model: Qwen/Qwen2.5-1.5B-Instruct
- Script: script_baseline_accepted.py
- Official submission: accepted
- Public score:
- EM:
- chrF:
- Notes:
## submission_001_rank1_baseline
- Date: 2026-07-08
- Score: 0.0855
- chrF: 0.147
- exact_match: 0.0498
- explanation_rate: 0
- Notes: First valid official submission.