Per-task downstream accuracy for mid-training ablation teacher_top1 using OLMo-2 7B Instruct teacher.
Source: source_paper:Table 9, Page 33
Reported and observed measurements
downstream_accuracy
t9_teacher_top1_gsm8k
Reported 62.2 percentage_points
Observed — percentage_points
downstream_accuracy
t9_teacher_top1_gsm_s
Reported 46 percentage_points
Observed — percentage_points
downstream_accuracy
t9_teacher_top1_gsmplus
Reported 40.1 percentage_points
Observed — percentage_points
downstream_accuracy
t9_teacher_top1_bbh
Reported 30.7 percentage_points
Observed — percentage_points
downstream_accuracy
t9_teacher_top1_drop
Reported 42.1 percentage_points
Observed — percentage_points
downstream_accuracy
t9_teacher_top1_math
Reported 8.8 percentage_points
Observed — percentage_points
downstream_accuracy
t9_teacher_top1_tqa
Reported 57.9 percentage_points
Observed — percentage_points
downstream_accuracy
t9_teacher_top1_nq
Reported 24.7 percentage_points
Observed — percentage_points
downstream_accuracy
t9_teacher_top1_sqa
Reported 9.3 percentage_points
Observed — percentage_points
downstream_accuracy
t9_teacher_top1_mmlu
Reported 48.8 percentage_points
Observed — percentage_points
downstream_accuracy
t9_teacher_top1_mmlu_p
Reported 18 percentage_points
Observed — percentage_points
downstream_accuracy
t9_teacher_top1_arc_c
Reported 59.8 percentage_points
Observed — percentage_points
downstream_accuracy
t9_teacher_top1_obqa
Reported 62 percentage_points
Observed — percentage_points
downstream_accuracy
t9_teacher_top1_wino
Reported 51 percentage_points
Observed — percentage_points
downstream_accuracy
t9_teacher_top1_agi
Reported 39.2 percentage_points
Observed — percentage_points
Assessments (0)
No immutable Assessment has been published for this Claim yet.
Runs (0)
No execution has been linked to this Claim yet.