CGA-bench/analysis/paper_runs/paper_fsm/task_summary.csv

1	model	condition	task_id	n_runs	coverage_mean	coverage_std	coverage_best	semantic_coverage_mean	semantic_coverage_std	semantic_coverage_best	eval2_mean	eval2_std	eval2_best	time_mean_sec	time_std_sec	token_cost_mean	token_cost_std	first_improvement_iter_mean
2	qwen-max	baseline	2012_q2fsm	1	91.66666666666666	0.0	91.66666666666666	61.17	0.0	61.17	0.7	0.0	0.7	348.97	0.0	0.62682	0.0
3	qwen-max	baseline	2013_q2afsm	1	92.3076923076923	0.0	92.3076923076923	73.51	0.0	73.51		0.0		519.09	0.0	0.7850400000000001	0.0
4	qwen-max	cga	2012_q2fsm	1	91.66666666666666	0.0	91.66666666666666	61.17	0.0	61.17	0.3	0.0	0.3	489.72	0.0	0.7213599999999999	0.0
5	qwen-max	cga	2013_q2afsm	1	92.3076923076923	0.0	92.3076923076923	73.51	0.0	73.51		0.0		1803.02	0.0	2.66142	0.0