Skip to content
GitLab
Menu
Projects
Groups
Snippets
Loading...
Help
Help
Support
Community forum
Keyboard shortcuts
?
Submit feedback
Contribute to GitLab
Sign in / Register
Toggle navigation
Menu
Open sidebar
gaoqiong
lm-evaluation-harness
Commits
b1ba4e71
Commit
b1ba4e71
authored
Oct 16, 2023
by
lintangsutawika
Browse files
added alt worlds prompts
parent
2afe7770
Changes
582
Hide whitespace changes
Inline
Side-by-side
Showing
20 changed files
with
91 additions
and
0 deletions
+91
-0
lm_eval/tasks/mmlu/alternative_worlds/full_continuation/style_01/style_01_marketing.yaml
...worlds/full_continuation/style_01/style_01_marketing.yaml
+4
-0
lm_eval/tasks/mmlu/alternative_worlds/full_continuation/style_01/style_01_medical_genetics.yaml
...full_continuation/style_01/style_01_medical_genetics.yaml
+4
-0
lm_eval/tasks/mmlu/alternative_worlds/full_continuation/style_01/style_01_miscellaneous.yaml
...ds/full_continuation/style_01/style_01_miscellaneous.yaml
+4
-0
lm_eval/tasks/mmlu/alternative_worlds/full_continuation/style_01/style_01_moral_disputes.yaml
...s/full_continuation/style_01/style_01_moral_disputes.yaml
+4
-0
lm_eval/tasks/mmlu/alternative_worlds/full_continuation/style_01/style_01_moral_scenarios.yaml
.../full_continuation/style_01/style_01_moral_scenarios.yaml
+4
-0
lm_eval/tasks/mmlu/alternative_worlds/full_continuation/style_01/style_01_nutrition.yaml
...worlds/full_continuation/style_01/style_01_nutrition.yaml
+4
-0
lm_eval/tasks/mmlu/alternative_worlds/full_continuation/style_01/style_01_philosophy.yaml
...orlds/full_continuation/style_01/style_01_philosophy.yaml
+4
-0
lm_eval/tasks/mmlu/alternative_worlds/full_continuation/style_01/style_01_prehistory.yaml
...orlds/full_continuation/style_01/style_01_prehistory.yaml
+4
-0
lm_eval/tasks/mmlu/alternative_worlds/full_continuation/style_01/style_01_professional_accounting.yaml
...ntinuation/style_01/style_01_professional_accounting.yaml
+4
-0
lm_eval/tasks/mmlu/alternative_worlds/full_continuation/style_01/style_01_professional_law.yaml
...full_continuation/style_01/style_01_professional_law.yaml
+4
-0
lm_eval/tasks/mmlu/alternative_worlds/full_continuation/style_01/style_01_professional_medicine.yaml
...continuation/style_01/style_01_professional_medicine.yaml
+4
-0
lm_eval/tasks/mmlu/alternative_worlds/full_continuation/style_01/style_01_professional_psychology.yaml
...ntinuation/style_01/style_01_professional_psychology.yaml
+4
-0
lm_eval/tasks/mmlu/alternative_worlds/full_continuation/style_01/style_01_public_relations.yaml
...full_continuation/style_01/style_01_public_relations.yaml
+4
-0
lm_eval/tasks/mmlu/alternative_worlds/full_continuation/style_01/style_01_security_studies.yaml
...full_continuation/style_01/style_01_security_studies.yaml
+4
-0
lm_eval/tasks/mmlu/alternative_worlds/full_continuation/style_01/style_01_sociology.yaml
...worlds/full_continuation/style_01/style_01_sociology.yaml
+4
-0
lm_eval/tasks/mmlu/alternative_worlds/full_continuation/style_01/style_01_us_foreign_policy.yaml
...ull_continuation/style_01/style_01_us_foreign_policy.yaml
+4
-0
lm_eval/tasks/mmlu/alternative_worlds/full_continuation/style_01/style_01_virology.yaml
..._worlds/full_continuation/style_01/style_01_virology.yaml
+4
-0
lm_eval/tasks/mmlu/alternative_worlds/full_continuation/style_01/style_01_world_religions.yaml
.../full_continuation/style_01/style_01_world_religions.yaml
+4
-0
lm_eval/tasks/mmlu/alternative_worlds/full_continuation/style_02/_template_yaml
...ernative_worlds/full_continuation/style_02/_template_yaml
+15
-0
lm_eval/tasks/mmlu/alternative_worlds/full_continuation/style_02/style_02_abstract_algebra.yaml
...full_continuation/style_02/style_02_abstract_algebra.yaml
+4
-0
No files found.
lm_eval/tasks/mmlu/alternative_worlds/full_continuation/style_01/style_01_marketing.yaml
0 → 100644
View file @
b1ba4e71
"
dataset_name"
:
"
marketing"
"
description"
:
"
The
following
are
multiple
choice
questions
(with
answers)
about
marketing.
\n\n
"
"
include"
:
"
_template_yaml"
"
task"
:
"
mmlu_style_01_fc_marketing"
lm_eval/tasks/mmlu/alternative_worlds/full_continuation/style_01/style_01_medical_genetics.yaml
0 → 100644
View file @
b1ba4e71
"
dataset_name"
:
"
medical_genetics"
"
description"
:
"
The
following
are
multiple
choice
questions
(with
answers)
about
medical
genetics.
\n\n
"
"
include"
:
"
_template_yaml"
"
task"
:
"
mmlu_style_01_fc_medical_genetics"
lm_eval/tasks/mmlu/alternative_worlds/full_continuation/style_01/style_01_miscellaneous.yaml
0 → 100644
View file @
b1ba4e71
"
dataset_name"
:
"
miscellaneous"
"
description"
:
"
The
following
are
multiple
choice
questions
(with
answers)
about
miscellaneous.
\n\n
"
"
include"
:
"
_template_yaml"
"
task"
:
"
mmlu_style_01_fc_miscellaneous"
lm_eval/tasks/mmlu/alternative_worlds/full_continuation/style_01/style_01_moral_disputes.yaml
0 → 100644
View file @
b1ba4e71
"
dataset_name"
:
"
moral_disputes"
"
description"
:
"
The
following
are
multiple
choice
questions
(with
answers)
about
moral
disputes.
\n\n
"
"
include"
:
"
_template_yaml"
"
task"
:
"
mmlu_style_01_fc_moral_disputes"
lm_eval/tasks/mmlu/alternative_worlds/full_continuation/style_01/style_01_moral_scenarios.yaml
0 → 100644
View file @
b1ba4e71
"
dataset_name"
:
"
moral_scenarios"
"
description"
:
"
The
following
are
multiple
choice
questions
(with
answers)
about
moral
scenarios.
\n\n
"
"
include"
:
"
_template_yaml"
"
task"
:
"
mmlu_style_01_fc_moral_scenarios"
lm_eval/tasks/mmlu/alternative_worlds/full_continuation/style_01/style_01_nutrition.yaml
0 → 100644
View file @
b1ba4e71
"
dataset_name"
:
"
nutrition"
"
description"
:
"
The
following
are
multiple
choice
questions
(with
answers)
about
nutrition.
\n\n
"
"
include"
:
"
_template_yaml"
"
task"
:
"
mmlu_style_01_fc_nutrition"
lm_eval/tasks/mmlu/alternative_worlds/full_continuation/style_01/style_01_philosophy.yaml
0 → 100644
View file @
b1ba4e71
"
dataset_name"
:
"
philosophy"
"
description"
:
"
The
following
are
multiple
choice
questions
(with
answers)
about
philosophy.
\n\n
"
"
include"
:
"
_template_yaml"
"
task"
:
"
mmlu_style_01_fc_philosophy"
lm_eval/tasks/mmlu/alternative_worlds/full_continuation/style_01/style_01_prehistory.yaml
0 → 100644
View file @
b1ba4e71
"
dataset_name"
:
"
prehistory"
"
description"
:
"
The
following
are
multiple
choice
questions
(with
answers)
about
prehistory.
\n\n
"
"
include"
:
"
_template_yaml"
"
task"
:
"
mmlu_style_01_fc_prehistory"
lm_eval/tasks/mmlu/alternative_worlds/full_continuation/style_01/style_01_professional_accounting.yaml
0 → 100644
View file @
b1ba4e71
"
dataset_name"
:
"
professional_accounting"
"
description"
:
"
The
following
are
multiple
choice
questions
(with
answers)
about
professional
accounting.
\n\n
"
"
include"
:
"
_template_yaml"
"
task"
:
"
mmlu_style_01_fc_professional_accounting"
lm_eval/tasks/mmlu/alternative_worlds/full_continuation/style_01/style_01_professional_law.yaml
0 → 100644
View file @
b1ba4e71
"
dataset_name"
:
"
professional_law"
"
description"
:
"
The
following
are
multiple
choice
questions
(with
answers)
about
professional
law.
\n\n
"
"
include"
:
"
_template_yaml"
"
task"
:
"
mmlu_style_01_fc_professional_law"
lm_eval/tasks/mmlu/alternative_worlds/full_continuation/style_01/style_01_professional_medicine.yaml
0 → 100644
View file @
b1ba4e71
"
dataset_name"
:
"
professional_medicine"
"
description"
:
"
The
following
are
multiple
choice
questions
(with
answers)
about
professional
medicine.
\n\n
"
"
include"
:
"
_template_yaml"
"
task"
:
"
mmlu_style_01_fc_professional_medicine"
lm_eval/tasks/mmlu/alternative_worlds/full_continuation/style_01/style_01_professional_psychology.yaml
0 → 100644
View file @
b1ba4e71
"
dataset_name"
:
"
professional_psychology"
"
description"
:
"
The
following
are
multiple
choice
questions
(with
answers)
about
professional
psychology.
\n\n
"
"
include"
:
"
_template_yaml"
"
task"
:
"
mmlu_style_01_fc_professional_psychology"
lm_eval/tasks/mmlu/alternative_worlds/full_continuation/style_01/style_01_public_relations.yaml
0 → 100644
View file @
b1ba4e71
"
dataset_name"
:
"
public_relations"
"
description"
:
"
The
following
are
multiple
choice
questions
(with
answers)
about
public
relations.
\n\n
"
"
include"
:
"
_template_yaml"
"
task"
:
"
mmlu_style_01_fc_public_relations"
lm_eval/tasks/mmlu/alternative_worlds/full_continuation/style_01/style_01_security_studies.yaml
0 → 100644
View file @
b1ba4e71
"
dataset_name"
:
"
security_studies"
"
description"
:
"
The
following
are
multiple
choice
questions
(with
answers)
about
security
studies.
\n\n
"
"
include"
:
"
_template_yaml"
"
task"
:
"
mmlu_style_01_fc_security_studies"
lm_eval/tasks/mmlu/alternative_worlds/full_continuation/style_01/style_01_sociology.yaml
0 → 100644
View file @
b1ba4e71
"
dataset_name"
:
"
sociology"
"
description"
:
"
The
following
are
multiple
choice
questions
(with
answers)
about
sociology.
\n\n
"
"
include"
:
"
_template_yaml"
"
task"
:
"
mmlu_style_01_fc_sociology"
lm_eval/tasks/mmlu/alternative_worlds/full_continuation/style_01/style_01_us_foreign_policy.yaml
0 → 100644
View file @
b1ba4e71
"
dataset_name"
:
"
us_foreign_policy"
"
description"
:
"
The
following
are
multiple
choice
questions
(with
answers)
about
us
foreign
policy.
\n\n
"
"
include"
:
"
_template_yaml"
"
task"
:
"
mmlu_style_01_fc_us_foreign_policy"
lm_eval/tasks/mmlu/alternative_worlds/full_continuation/style_01/style_01_virology.yaml
0 → 100644
View file @
b1ba4e71
"
dataset_name"
:
"
virology"
"
description"
:
"
The
following
are
multiple
choice
questions
(with
answers)
about
virology.
\n\n
"
"
include"
:
"
_template_yaml"
"
task"
:
"
mmlu_style_01_fc_virology"
lm_eval/tasks/mmlu/alternative_worlds/full_continuation/style_01/style_01_world_religions.yaml
0 → 100644
View file @
b1ba4e71
"
dataset_name"
:
"
world_religions"
"
description"
:
"
The
following
are
multiple
choice
questions
(with
answers)
about
world
religions.
\n\n
"
"
include"
:
"
_template_yaml"
"
task"
:
"
mmlu_style_01_fc_world_religions"
lm_eval/tasks/mmlu/alternative_worlds/full_continuation/style_02/_template_yaml
0 → 100644
View file @
b1ba4e71
group: mmlu_style_02_fc
dataset_path: cais/mmlu
test_split: test
fewshot_split: dev
output_type: multiple_choice
doc_to_text: "Q: {{question.strip()}}\nA. {{choices[0]}}\nB. {{choices[1]}}\nC. {{choices[2]}}\nD. {{choices[3]}}\nA: "
doc_to_choice: "{{['A. '+choices[0], 'B. '+choices[1], 'C. '+choices[2], 'D. '+choices[3]]}}"
doc_to_target: answer
metric_list:
- metric: acc
aggregation: mean
higher_is_better: true
- metric: acc_norm
aggregation: mean
higher_is_better: true
lm_eval/tasks/mmlu/alternative_worlds/full_continuation/style_02/style_02_abstract_algebra.yaml
0 → 100644
View file @
b1ba4e71
"
dataset_name"
:
"
abstract_algebra"
"
description"
:
"
The
following
are
multiple
choice
questions
(with
answers)
about
abstract
algebra.
\n\n
"
"
include"
:
"
_template_yaml"
"
task"
:
"
mmlu_style_02_fc_abstract_algebra"
Prev
1
2
3
4
5
6
7
…
30
Next
Write
Preview
Markdown
is supported
0%
Try again
or
attach a new file
.
Attach a file
Cancel
You are about to add
0
people
to the discussion. Proceed with caution.
Finish editing this message first!
Cancel
Please
register
or
sign in
to comment