Downstream task evaluation after post-training stage RLVR1 for 13b_fkd.
Source: source_paper:Table 8, Page 33, After RLVR1
Reported and observed measurements
downstream_accuracy
t8_rlvr1_13b_fkd_gsm8k
Reported 73 percentage_points
Observed — percentage_points
downstream_accuracy
t8_rlvr1_13b_fkd_gsm_s
Reported 55.5 percentage_points
Observed — percentage_points
downstream_accuracy
t8_rlvr1_13b_fkd_gsmplus
Reported 47.3 percentage_points
Observed — percentage_points
downstream_accuracy
t8_rlvr1_13b_fkd_bbh
Reported 33.7 percentage_points
Observed — percentage_points
downstream_accuracy
t8_rlvr1_13b_fkd_drop
Reported 36.1 percentage_points
Observed — percentage_points
downstream_accuracy
t8_rlvr1_13b_fkd_math
Reported 14 percentage_points
Observed — percentage_points
downstream_accuracy
t8_rlvr1_13b_fkd_tqa
Reported 54.7 percentage_points
Observed — percentage_points
downstream_accuracy
t8_rlvr1_13b_fkd_nq
Reported 23.3 percentage_points
Observed — percentage_points
downstream_accuracy
t8_rlvr1_13b_fkd_sqa
Reported 8.3 percentage_points
Observed — percentage_points
downstream_accuracy
t8_rlvr1_13b_fkd_mmlu
Reported 35 percentage_points
Observed — percentage_points
downstream_accuracy
t8_rlvr1_13b_fkd_mmlu_p
Reported 17.2 percentage_points
Observed — percentage_points
downstream_accuracy
t8_rlvr1_13b_fkd_arc_c
Reported 54 percentage_points
Observed — percentage_points
downstream_accuracy
t8_rlvr1_13b_fkd_obqa
Reported 56.6 percentage_points
Observed — percentage_points
downstream_accuracy
t8_rlvr1_13b_fkd_wino
Reported 51.7 percentage_points
Observed — percentage_points
downstream_accuracy
t8_rlvr1_13b_fkd_agi
Reported 35.4 percentage_points
Observed — percentage_points
downstream_accuracy
t8_rlvr1_13b_fkd_ife
Reported 63.2 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.