README.md

mlfoundations-dev/a1_code_code_golf_eval_636d

Precomputed model outputs for evaluation.

Evaluation Results

Summary

Metric	AIME24	AMC23	MATH500	MMLUPro	JEEBench	GPQADiamond	LiveCodeBench	CodeElo	CodeForces
Accuracy	17.7	58.0	77.0	30.6	38.8	42.6	44.4	14.4	17.2

AIME24

• Average Accuracy: 17.67% ± 1.95%
• Number of Runs: 10

Run	Accuracy	Questions Solved	Total Questions
1	10.00%	3	30
2	26.67%	8	30
3	30.00%	9	30
4	16.67%	5	30
5	13.33%	4	30
6	10.00%	3	30
7	16.67%	5	30
8	16.67%	5	30
9	16.67%	5	30
10	20.00%	6	30

AMC23

• Average Accuracy: 58.00% ± 0.92%
• Number of Runs: 10

Run	Accuracy	Questions Solved	Total Questions
1	60.00%	24	40
2	57.50%	23	40
3	55.00%	22	40
4	52.50%	21	40
5	62.50%	25	40
6	57.50%	23	40
7	60.00%	24	40
8	55.00%	22	40
9	60.00%	24	40
10	60.00%	24	40

MATH500

• Accuracy: 77.00%

  Accuracy	Questions Solved	Total Questions
  77.00%	385	500

MMLUPro

• Average Accuracy: 30.60% ± 0.00%
• Number of Runs: 1

Run	Accuracy	Questions Solved	Total Questions
1	30.60%	153	500

JEEBench

• Average Accuracy: 38.83% ± 0.77%
• Number of Runs: 3

Run	Accuracy	Questions Solved	Total Questions
1	36.99%	190.5	515
2	40.05%	206.25	515
3	39.47%	203.25	515

GPQADiamond

• Average Accuracy: 42.59% ± 0.60%
• Number of Runs: 3

Run	Accuracy	Questions Solved	Total Questions
1	41.41%	82	198
2	42.42%	84	198
3	43.94%	87	198

LiveCodeBench

• Average Accuracy: 44.42% ± 1.27%
• Number of Runs: 3

Run	Accuracy	Questions Solved	Total Questions
1	45.79%	234	511
2	45.60%	233	511
3	41.88%	214	511

CodeElo

• Average Accuracy: 14.41% ± 0.90%
• Number of Runs: 3

Run	Accuracy	Questions Solved	Total Questions
1	14.07%	55	391
2	13.04%	51	391
3	16.11%	63	391

CodeForces

• Average Accuracy: 17.22% ± 0.80%
• Number of Runs: 3

Run	Accuracy	Questions Solved	Total Questions
1	16.11%	73	453
2	16.78%	76	453
3	18.76%	85	453