Downstream task evaluation after post-training stage RLVR1 for 13b_sd.
Source: source_paper:Table 8, Page 33, After RLVR1
Reported and observed measurements
downstream_accuracy
t8_rlvr1_13b_sd_gsm8k
Reported 78.8 percentage_points
Observed — percentage_points
downstream_accuracy
t8_rlvr1_13b_sd_gsm_s
Reported 65.6 percentage_points
Observed — percentage_points
downstream_accuracy
t8_rlvr1_13b_sd_gsmplus
Reported 52.5 percentage_points
Observed — percentage_points
downstream_accuracy
t8_rlvr1_13b_sd_bbh
Reported 33.5 percentage_points
Observed — percentage_points
downstream_accuracy
t8_rlvr1_13b_sd_drop
Reported 43.3 percentage_points
Observed — percentage_points
downstream_accuracy
t8_rlvr1_13b_sd_math
Reported 17.8 percentage_points
Observed — percentage_points
downstream_accuracy
t8_rlvr1_13b_sd_tqa
Reported 54.3 percentage_points
Observed — percentage_points
downstream_accuracy
t8_rlvr1_13b_sd_nq
Reported 24 percentage_points
Observed — percentage_points
downstream_accuracy
t8_rlvr1_13b_sd_sqa
Reported 8.2 percentage_points
Observed — percentage_points
downstream_accuracy
t8_rlvr1_13b_sd_mmlu
Reported 38.7 percentage_points
Observed — percentage_points
downstream_accuracy
t8_rlvr1_13b_sd_mmlu_p
Reported 17.7 percentage_points
Observed — percentage_points
downstream_accuracy
t8_rlvr1_13b_sd_arc_c
Reported 56.2 percentage_points
Observed — percentage_points
downstream_accuracy
t8_rlvr1_13b_sd_obqa
Reported 57.8 percentage_points
Observed — percentage_points
downstream_accuracy
t8_rlvr1_13b_sd_wino
Reported 51.7 percentage_points
Observed — percentage_points
downstream_accuracy
t8_rlvr1_13b_sd_agi
Reported 37.7 percentage_points
Observed — percentage_points
downstream_accuracy
t8_rlvr1_13b_sd_ife
Reported 68.9 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.