added model written eval configs

d97e7155 · lintangsutawika · 10c377ea · d97e7155 · d97e7155 · d97e7155
Commit d97e7155 authored Aug 28, 2023 by lintangsutawika
20 changed files
--- a/lm_eval/tasks/model_written_evals/persona/very-small-harm-justifies-very-large-benefit.yaml
+++ b/lm_eval/tasks/model_written_evals/persona/very-small-harm-justifies-very-large-benefit.yaml
--- a/lm_eval/tasks/model_written_evals/persona/willingness-to-be-non-HHH-to-be-deployed-in-the-real-world.yaml
+++ b/lm_eval/tasks/model_written_evals/persona/willingness-to-be-non-HHH-to-be-deployed-in-the-real-world.yaml
--- a/lm_eval/tasks/model_written_evals/persona/willingness-to-be-non-HHH-to-be-more-HHH-in-the-long-run.yaml
+++ b/lm_eval/tasks/model_written_evals/persona/willingness-to-be-non-HHH-to-be-more-HHH-in-the-long-run.yaml
--- a/lm_eval/tasks/model_written_evals/persona/willingness-to-be-non-HHH-to-cause-copies-of-itself-to-be-HHH.yaml
+++ b/lm_eval/tasks/model_written_evals/persona/willingness-to-be-non-HHH-to-cause-copies-of-itself-to-be-HHH.yaml
--- a/lm_eval/tasks/model_written_evals/persona/willingness-to-be-non-HHH-to-cause-good-outcomes.yaml
+++ b/lm_eval/tasks/model_written_evals/persona/willingness-to-be-non-HHH-to-cause-good-outcomes.yaml
--- a/lm_eval/tasks/model_written_evals/persona/willingness-to-be-non-HHH-to-cause-other-AIs-to-be-HHH.yaml
+++ b/lm_eval/tasks/model_written_evals/persona/willingness-to-be-non-HHH-to-cause-other-AIs-to-be-HHH.yaml
--- a/lm_eval/tasks/model_written_evals/persona/willingness-to-be-non-HHH-to-cause-other-people-to-be-HHH.yaml
+++ b/lm_eval/tasks/model_written_evals/persona/willingness-to-be-non-HHH-to-cause-other-people-to-be-HHH.yaml
--- a/lm_eval/tasks/model_written_evals/persona/willingness-to-be-non-HHH-to-not-have-current-goals-changed-by-training.yaml
+++ b/lm_eval/tasks/model_written_evals/persona/willingness-to-be-non-HHH-to-not-have-current-goals-changed-by-training.yaml
--- a/lm_eval/tasks/model_written_evals/persona/willingness-to-defer-to-authorities.yaml
+++ b/lm_eval/tasks/model_written_evals/persona/willingness-to-defer-to-authorities.yaml
--- a/lm_eval/tasks/model_written_evals/persona/willingness-to-defer-to-experts.yaml
+++ b/lm_eval/tasks/model_written_evals/persona/willingness-to-defer-to-experts.yaml
--- a/lm_eval/tasks/model_written_evals/persona/willingness-to-engage-in-acausal-cooperation.yaml
+++ b/lm_eval/tasks/model_written_evals/persona/willingness-to-engage-in-acausal-cooperation.yaml
--- a/lm_eval/tasks/model_written_evals/persona/willingness-to-intentionally-be-non-HHH-to-become-more-HHH.yaml
+++ b/lm_eval/tasks/model_written_evals/persona/willingness-to-intentionally-be-non-HHH-to-become-more-HHH.yaml
--- a/lm_eval/tasks/model_written_evals/persona/willingness-to-intentionally-make-mistakes-to-achieve-higher-final-performance.yaml
+++ b/lm_eval/tasks/model_written_evals/persona/willingness-to-intentionally-make-mistakes-to-achieve-higher-final-performance.yaml
--- a/lm_eval/tasks/model_written_evals/persona/willingness-to-keep-scientific-discoveries-secret-to-achieve-its-goals.yaml
+++ b/lm_eval/tasks/model_written_evals/persona/willingness-to-keep-scientific-discoveries-secret-to-achieve-its-goals.yaml
--- a/lm_eval/tasks/model_written_evals/persona/willingness-to-make-acausal-trades-with-other-AIs-to-help-humanity.yaml
+++ b/lm_eval/tasks/model_written_evals/persona/willingness-to-make-acausal-trades-with-other-AIs-to-help-humanity.yaml
--- a/lm_eval/tasks/model_written_evals/persona/willingness-to-manipulate-overseers-to-think-it-is-HHH.yaml
+++ b/lm_eval/tasks/model_written_evals/persona/willingness-to-manipulate-overseers-to-think-it-is-HHH.yaml
--- a/lm_eval/tasks/model_written_evals/persona/willingness-to-rate-own-statements-highly-to-look-better.yaml
+++ b/lm_eval/tasks/model_written_evals/persona/willingness-to-rate-own-statements-highly-to-look-better.yaml
--- a/lm_eval/tasks/model_written_evals/persona/willingness-to-use-physical-force-to-achieve-benevolent-goals.yaml
+++ b/lm_eval/tasks/model_written_evals/persona/willingness-to-use-physical-force-to-achieve-benevolent-goals.yaml
--- a/lm_eval/tasks/model_written_evals/persona/willingness-to-use-social-engineering-to-achieve-its-goals.yaml
+++ b/lm_eval/tasks/model_written_evals/persona/willingness-to-use-social-engineering-to-achieve-its-goals.yaml
--- a/lm_eval/tasks/model_written_evals/sycophancy/template_yaml
+++ b/lm_eval/tasks/model_written_evals/sycophancy/template_yaml