Downstream task evaluation after post-training stage RLVR1 for ntp_shared.
Source: source_paper:Table 8, Page 33, After RLVR1
Reported and observed measurements
downstream_accuracy
t8_rlvr1_ntp_shared_gsm8k
Reported 69.1 percentage_points
Observed — percentage_points
downstream_accuracy
t8_rlvr1_ntp_shared_gsm_s
Reported 45.6 percentage_points
Observed — percentage_points
downstream_accuracy
t8_rlvr1_ntp_shared_gsmplus
Reported 40.5 percentage_points
Observed — percentage_points
downstream_accuracy
t8_rlvr1_ntp_shared_bbh
Reported 32.4 percentage_points
Observed — percentage_points
downstream_accuracy
t8_rlvr1_ntp_shared_drop
Reported 33.4 percentage_points
Observed — percentage_points
downstream_accuracy
t8_rlvr1_ntp_shared_math
Reported 11.2 percentage_points
Observed — percentage_points
downstream_accuracy
t8_rlvr1_ntp_shared_tqa
Reported 54.5 percentage_points
Observed — percentage_points
downstream_accuracy
t8_rlvr1_ntp_shared_nq
Reported 21.7 percentage_points
Observed — percentage_points
downstream_accuracy
t8_rlvr1_ntp_shared_sqa
Reported 8 percentage_points
Observed — percentage_points
downstream_accuracy
t8_rlvr1_ntp_shared_mmlu
Reported 31.1 percentage_points
Observed — percentage_points
downstream_accuracy
t8_rlvr1_ntp_shared_mmlu_p
Reported 16.9 percentage_points
Observed — percentage_points
downstream_accuracy
t8_rlvr1_ntp_shared_arc_c
Reported 52.3 percentage_points
Observed — percentage_points
downstream_accuracy
t8_rlvr1_ntp_shared_obqa
Reported 54.2 percentage_points
Observed — percentage_points
downstream_accuracy
t8_rlvr1_ntp_shared_wino
Reported 51.1 percentage_points
Observed — percentage_points
downstream_accuracy
t8_rlvr1_ntp_shared_agi
Reported 33.2 percentage_points
Observed — percentage_points
downstream_accuracy
t8_rlvr1_ntp_shared_ife
Reported 63.6 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.