README.md

mlfoundations-dev/a1_code_stackexchange_eval_636d

Precomputed model outputs for evaluation.

Evaluation Results

Summary

Metric	AIME24	AMC23	MATH500	MMLUPro	JEEBench	GPQADiamond	LiveCodeBench	CodeElo	CodeForces
Accuracy	14.7	54.2	73.6	28.6	36.1	38.6	12.0	5.6	7.9

AIME24

• Average Accuracy: 14.67% ± 1.17%
• Number of Runs: 10

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

AMC23

• Average Accuracy: 54.25% ± 1.66%
• Number of Runs: 10

Run	Accuracy	Questions Solved	Total Questions
1	47.50%	19	40
2	47.50%	19	40
3	65.00%	26	40
4	52.50%	21	40
5	55.00%	22	40
6	55.00%	22	40
7	52.50%	21	40
8	57.50%	23	40
9	50.00%	20	40
10	60.00%	24	40

MATH500

• Accuracy: 73.60%

  Accuracy	Questions Solved	Total Questions
  73.60%	368	500

MMLUPro

• Average Accuracy: 28.60% ± 0.00%
• Number of Runs: 1

Run	Accuracy	Questions Solved	Total Questions
1	28.60%	143	500

JEEBench

• Average Accuracy: 36.08% ± 0.56%
• Number of Runs: 3

Run	Accuracy	Questions Solved	Total Questions
1	35.05%	180.5	515
2	35.83%	184.5	515
3	37.38%	192.5	515

GPQADiamond

• Average Accuracy: 38.55% ± 0.14%
• Number of Runs: 3

Run	Accuracy	Questions Solved	Total Questions
1	38.89%	77	198
2	38.38%	76	198
3	38.38%	76	198

LiveCodeBench

• Average Accuracy: 12.00% ± 0.75%
• Number of Runs: 3

Run	Accuracy	Questions Solved	Total Questions
1	13.11%	67	511
2	10.57%	54	511
3	12.33%	63	511

CodeElo

• Average Accuracy: 5.63% ± 0.74%
• Number of Runs: 3

Run	Accuracy	Questions Solved	Total Questions
1	5.63%	22	391
2	6.91%	27	391
3	4.35%	17	391

CodeForces

• Average Accuracy: 7.87% ± 0.72%
• Number of Runs: 3

Run	Accuracy	Questions Solved	Total Questions
1	7.51%	34	453
2	9.27%	42	453
3	6.84%	31	453