Downstream task evaluation after post-training stage DPO for 13b_fkd.
Source: source_paper:Table 8, Page 33, After DPO
Reported and observed measurements
downstream_accuracy
t8_dpo_13b_fkd_gsm8k
Reported 61 percentage_points
Observed — percentage_points
downstream_accuracy
t8_dpo_13b_fkd_gsm_s
Reported 34.2 percentage_points
Observed — percentage_points
downstream_accuracy
t8_dpo_13b_fkd_gsmplus
Reported 31.7 percentage_points
Observed — percentage_points
downstream_accuracy
t8_dpo_13b_fkd_bbh
Reported 33.4 percentage_points
Observed — percentage_points
downstream_accuracy
t8_dpo_13b_fkd_drop
Reported 36.9 percentage_points
Observed — percentage_points
downstream_accuracy
t8_dpo_13b_fkd_math
Reported 8.2 percentage_points
Observed — percentage_points
downstream_accuracy
t8_dpo_13b_fkd_tqa
Reported 54.6 percentage_points
Observed — percentage_points
downstream_accuracy
t8_dpo_13b_fkd_nq
Reported 23.2 percentage_points
Observed — percentage_points
downstream_accuracy
t8_dpo_13b_fkd_sqa
Reported 7.7 percentage_points
Observed — percentage_points
downstream_accuracy
t8_dpo_13b_fkd_mmlu
Reported 45.9 percentage_points
Observed — percentage_points
downstream_accuracy
t8_dpo_13b_fkd_mmlu_p
Reported 17.4 percentage_points
Observed — percentage_points
downstream_accuracy
t8_dpo_13b_fkd_arc_c
Reported 53.2 percentage_points
Observed — percentage_points
downstream_accuracy
t8_dpo_13b_fkd_obqa
Reported 57.2 percentage_points
Observed — percentage_points
downstream_accuracy
t8_dpo_13b_fkd_wino
Reported 52.3 percentage_points
Observed — percentage_points
downstream_accuracy
t8_dpo_13b_fkd_agi
Reported 37.4 percentage_points
Observed — percentage_points
downstream_accuracy
t8_dpo_13b_fkd_ife
Reported 64.7 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.