Downstream task evaluation after post-training stage RLVR1 for 7b_rkd.
Source: source_paper:Table 8, Page 33, After RLVR1
Reported and observed measurements
downstream_accuracy
t8_rlvr1_7b_rkd_gsm8k
Reported 77.9 percentage_points
Observed — percentage_points
downstream_accuracy
t8_rlvr1_7b_rkd_gsm_s
Reported 54.3 percentage_points
Observed — percentage_points
downstream_accuracy
t8_rlvr1_7b_rkd_gsmplus
Reported 52.5 percentage_points
Observed — percentage_points
downstream_accuracy
t8_rlvr1_7b_rkd_bbh
Reported 34.2 percentage_points
Observed — percentage_points
downstream_accuracy
t8_rlvr1_7b_rkd_drop
Reported 40.1 percentage_points
Observed — percentage_points
downstream_accuracy
t8_rlvr1_7b_rkd_math
Reported 15.6 percentage_points
Observed — percentage_points
downstream_accuracy
t8_rlvr1_7b_rkd_tqa
Reported 52.1 percentage_points
Observed — percentage_points
downstream_accuracy
t8_rlvr1_7b_rkd_nq
Reported 21.9 percentage_points
Observed — percentage_points
downstream_accuracy
t8_rlvr1_7b_rkd_sqa
Reported 7.8 percentage_points
Observed — percentage_points
downstream_accuracy
t8_rlvr1_7b_rkd_mmlu
Reported 45.3 percentage_points
Observed — percentage_points
downstream_accuracy
t8_rlvr1_7b_rkd_mmlu_p
Reported 16.8 percentage_points
Observed — percentage_points
downstream_accuracy
t8_rlvr1_7b_rkd_arc_c
Reported 57.6 percentage_points
Observed — percentage_points
downstream_accuracy
t8_rlvr1_7b_rkd_obqa
Reported 50.4 percentage_points
Observed — percentage_points
downstream_accuracy
t8_rlvr1_7b_rkd_wino
Reported 51.5 percentage_points
Observed — percentage_points
downstream_accuracy
t8_rlvr1_7b_rkd_agi
Reported 36.1 percentage_points
Observed — percentage_points
downstream_accuracy
t8_rlvr1_7b_rkd_ife
Reported 59.3 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.