--- license: mit tags: - completions - vllm - chemistry --- # wtd/rewards-chemistry-olmo-3-7b-instruct Per-question data collected with **allenai/Olmo-3-7B-Instruct** on the **chemistry** dataset. ## Generation schedule | Temperature | # completions | |-------------|---------------| | 1.0 | 64 | ## Reward mean histogram | Bin | Count | |-----|-------| | 0.0–0.1 | █████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████ (145) | | 0.1–0.2 | █████████████████████████████████████████████████████████ (57) | | 0.2–0.3 | ██████████████████████████████████████████████████████████████████████████ (74) | | 0.3–0.4 | ████████████████████████████████████████████████ (48) | | 0.4–0.5 | ███████████████████████████████████████ (39) | | 0.5–0.6 | █████████████████████████████████████████ (41) | | 0.6–0.7 | █████████████████████████ (25) | | 0.7–0.8 | ██████████████████████████████████ (34) | | 0.8–0.9 | ███████████████████████ (23) | | 0.9–1.0 | █████████████████ (17) | | 1.0–1.1 | █████████ (9) | ## Schema - `reward` (`dict`): {"scorer": str, "rewards": list[float], "reward_mean": float}. - `model_id` (`str`): model used for generation. - `dataset_name` (`str`): source dataset (`omni_math`, `chemistry`, or `reward_bench`). - All original question fields from the source dataset are preserved.