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
91b2eec6
"tests/ds_config_bf16.json" did not exist on "8ec5d6780e546aaa6338b2d3271f291d4ecc3127"
Commit
91b2eec6
authored
Jun 13, 2024
by
Yu Shi Jie
Browse files
added generative mmlu-pro
parent
6c5edc99
Changes
17
Hide whitespace changes
Inline
Side-by-side
Showing
17 changed files
with
140 additions
and
4 deletions
+140
-4
lm_eval/tasks/mmlu_pro/default/mmlu_pro_history.yaml
lm_eval/tasks/mmlu_pro/default/mmlu_pro_history.yaml
+4
-4
lm_eval/tasks/mmlu_pro/generative/_default_template_yaml
lm_eval/tasks/mmlu_pro/generative/_default_template_yaml
+18
-0
lm_eval/tasks/mmlu_pro/generative/_mmlu_pro.yaml
lm_eval/tasks/mmlu_pro/generative/_mmlu_pro.yaml
+6
-0
lm_eval/tasks/mmlu_pro/generative/mmlu_pro_biology.yaml
lm_eval/tasks/mmlu_pro/generative/mmlu_pro_biology.yaml
+8
-0
lm_eval/tasks/mmlu_pro/generative/mmlu_pro_business.yaml
lm_eval/tasks/mmlu_pro/generative/mmlu_pro_business.yaml
+8
-0
lm_eval/tasks/mmlu_pro/generative/mmlu_pro_chemistry.yaml
lm_eval/tasks/mmlu_pro/generative/mmlu_pro_chemistry.yaml
+8
-0
lm_eval/tasks/mmlu_pro/generative/mmlu_pro_computer_science.yaml
.../tasks/mmlu_pro/generative/mmlu_pro_computer_science.yaml
+8
-0
lm_eval/tasks/mmlu_pro/generative/mmlu_pro_economics.yaml
lm_eval/tasks/mmlu_pro/generative/mmlu_pro_economics.yaml
+8
-0
lm_eval/tasks/mmlu_pro/generative/mmlu_pro_engineering.yaml
lm_eval/tasks/mmlu_pro/generative/mmlu_pro_engineering.yaml
+8
-0
lm_eval/tasks/mmlu_pro/generative/mmlu_pro_health.yaml
lm_eval/tasks/mmlu_pro/generative/mmlu_pro_health.yaml
+8
-0
lm_eval/tasks/mmlu_pro/generative/mmlu_pro_history.yaml
lm_eval/tasks/mmlu_pro/generative/mmlu_pro_history.yaml
+8
-0
lm_eval/tasks/mmlu_pro/generative/mmlu_pro_law.yaml
lm_eval/tasks/mmlu_pro/generative/mmlu_pro_law.yaml
+8
-0
lm_eval/tasks/mmlu_pro/generative/mmlu_pro_math.yaml
lm_eval/tasks/mmlu_pro/generative/mmlu_pro_math.yaml
+8
-0
lm_eval/tasks/mmlu_pro/generative/mmlu_pro_other.yaml
lm_eval/tasks/mmlu_pro/generative/mmlu_pro_other.yaml
+8
-0
lm_eval/tasks/mmlu_pro/generative/mmlu_pro_philosophy.yaml
lm_eval/tasks/mmlu_pro/generative/mmlu_pro_philosophy.yaml
+8
-0
lm_eval/tasks/mmlu_pro/generative/mmlu_pro_physics.yaml
lm_eval/tasks/mmlu_pro/generative/mmlu_pro_physics.yaml
+8
-0
lm_eval/tasks/mmlu_pro/generative/mmlu_pro_psychology.yaml
lm_eval/tasks/mmlu_pro/generative/mmlu_pro_psychology.yaml
+8
-0
No files found.
lm_eval/tasks/mmlu_pro/default/mmlu_pro_history.yaml
View file @
91b2eec6
"
dataset_name"
:
"
law
"
"
description"
:
"
The
following
are
multiple
choice
questions
(with
answers)
about
law
.
\n\
"
dataset_name"
:
"
history
"
"
description"
:
"
The
following
are
multiple
choice
questions
(with
answers)
about
history
.
\n\
\n
"
"
group"
:
"
mmlu_pro_humanities"
"
group_alias"
:
"
humanities"
"
include"
:
"
_default_template_yaml"
"
task"
:
"
mmlu_pro_
law
"
"
task_alias"
:
"
law
"
"
task"
:
"
mmlu_pro_
history
"
"
task_alias"
:
"
history
"
lm_eval/tasks/mmlu_pro/generative/_default_template_yaml
0 → 100644
View file @
91b2eec6
dataset_path: sjyuxyz/MMLU-Pro-with-subset
test_split: test
fewshot_split: validation
fewshot_config:
sampler: first_n
output_type: generate_until
doc_to_text: "{{question.strip()}}\nA. {{choices[0]}}\nB. {{choices[1]}}\nC. {{choices[2]}}\nD. {{choices[3]}}\nE. {{choices[4]}}\nF. {{choices[5]}}\nG. {{choices[6]}}\nH. {{choices[7]}}\nI. {{choices[8]}}\nJ. {{choices[9]}}\nAnswer:"
doc_to_target: "{{['A', 'B', 'C', 'D', 'E', 'F', 'G', 'H', 'I', 'J'][answer]}}"
generation_kwargs:
until:
- "</s>"
- "\n"
metric_list:
- metric: exact_match
aggregation: mean
higher_is_better: true
metadata:
version: 1.0
lm_eval/tasks/mmlu_pro/generative/_mmlu_pro.yaml
0 → 100644
View file @
91b2eec6
group
:
mmlu_pro_generative
task
:
-
mmlu_pro_stem_generative
-
mmlu_pro_other_generative
-
mmlu_pro_social_sciences_generative
-
mmlu_pro_humanities_generative
lm_eval/tasks/mmlu_pro/
default/mmlu_abstract_algebra
.yaml
→
lm_eval/tasks/mmlu_pro/
generative/mmlu_pro_biology
.yaml
View file @
91b2eec6
"
dataset_name"
:
"
abstract_algebra
"
"
description"
:
"
The
following
are
multiple
choice
questions
(with
answers)
about
abstract
\
\
algebra.
\n\
n
"
"
group"
:
"
mmlu_
stem
"
"
dataset_name"
:
"
biology
"
"
description"
:
"
The
following
are
multiple
choice
questions
(with
answers)
about
biology.
\n
\
\n
"
"
group"
:
"
mmlu_
pro_stem_generative
"
"
group_alias"
:
"
stem"
"
include"
:
"
_default_template_yaml"
"
task"
:
"
mmlu_
abstract_algebra
"
"
task_alias"
:
"
abstract_algebra
"
"
task"
:
"
mmlu_
pro_biology_generative
"
"
task_alias"
:
"
biology
"
lm_eval/tasks/mmlu_pro/generative/mmlu_pro_business.yaml
0 → 100644
View file @
91b2eec6
"
dataset_name"
:
"
business"
"
description"
:
"
The
following
are
multiple
choice
questions
(with
answers)
about
business.
\n\
\n
"
"
group"
:
"
mmlu_pro_other_generative"
"
group_alias"
:
"
other"
"
include"
:
"
_default_template_yaml"
"
task"
:
"
mmlu_pro_business_generative"
"
task_alias"
:
"
business"
lm_eval/tasks/mmlu_pro/
default/mmlu_anatom
y.yaml
→
lm_eval/tasks/mmlu_pro/
generative/mmlu_pro_chemistr
y.yaml
View file @
91b2eec6
"
dataset_name"
:
"
anatomy
"
"
description"
:
"
The
following
are
multiple
choice
questions
(with
answers)
about
anatomy
.
\n\
"
dataset_name"
:
"
math
"
"
description"
:
"
The
following
are
multiple
choice
questions
(with
answers)
about
math
.
\n\
\n
"
"
group"
:
"
mmlu_
stem
"
"
group"
:
"
mmlu_
pro_stem_generative
"
"
group_alias"
:
"
stem"
"
include"
:
"
_default_template_yaml"
"
task"
:
"
mmlu_
anatomy
"
"
task_alias"
:
"
anatomy
"
"
task"
:
"
mmlu_
pro_math_generative
"
"
task_alias"
:
"
math
"
lm_eval/tasks/mmlu_pro/generative/mmlu_pro_computer_science.yaml
0 → 100644
View file @
91b2eec6
"
dataset_name"
:
"
computer_science"
"
description"
:
"
The
following
are
multiple
choice
questions
(with
answers)
about
computer_science.
\n\
\n
"
"
group"
:
"
mmlu_pro_stem_generative"
"
group_alias"
:
"
stem"
"
include"
:
"
_default_template_yaml"
"
task"
:
"
mmlu_pro_computer_science_generative"
"
task_alias"
:
"
computer_science"
lm_eval/tasks/mmlu_pro/generative/mmlu_pro_economics.yaml
0 → 100644
View file @
91b2eec6
"
dataset_name"
:
"
economics"
"
description"
:
"
The
following
are
multiple
choice
questions
(with
answers)
about
economics.
\n\
\n
"
"
group"
:
"
mmlu_pro_social_sciences_generative"
"
group_alias"
:
"
social_sciences"
"
include"
:
"
_default_template_yaml"
"
task"
:
"
mmlu_pro_economics_generative"
"
task_alias"
:
"
economics"
lm_eval/tasks/mmlu_pro/generative/mmlu_pro_engineering.yaml
0 → 100644
View file @
91b2eec6
"
dataset_name"
:
"
engineering"
"
description"
:
"
The
following
are
multiple
choice
questions
(with
answers)
about
engineering.
\n\
\n
"
"
group"
:
"
mmlu_pro_stem_generative"
"
group_alias"
:
"
stem"
"
include"
:
"
_default_template_yaml"
"
task"
:
"
mmlu_pro_engineering_generative"
"
task_alias"
:
"
engineering"
lm_eval/tasks/mmlu_pro/generative/mmlu_pro_health.yaml
0 → 100644
View file @
91b2eec6
"
dataset_name"
:
"
health"
"
description"
:
"
The
following
are
multiple
choice
questions
(with
answers)
about
health.
\n\
\n
"
"
group"
:
"
mmlu_pro_other_generative"
"
group_alias"
:
"
other"
"
include"
:
"
_default_template_yaml"
"
task"
:
"
mmlu_pro_health_generative"
"
task_alias"
:
"
health"
lm_eval/tasks/mmlu_pro/generative/mmlu_pro_history.yaml
0 → 100644
View file @
91b2eec6
"
dataset_name"
:
"
history"
"
description"
:
"
The
following
are
multiple
choice
questions
(with
answers)
about
history.
\n\
\n
"
"
group"
:
"
mmlu_pro_humanities_generative"
"
group_alias"
:
"
humanities"
"
include"
:
"
_default_template_yaml"
"
task"
:
"
mmlu_pro_history_generative"
"
task_alias"
:
"
history"
lm_eval/tasks/mmlu_pro/generative/mmlu_pro_law.yaml
0 → 100644
View file @
91b2eec6
"
dataset_name"
:
"
law"
"
description"
:
"
The
following
are
multiple
choice
questions
(with
answers)
about
law.
\n\
\n
"
"
group"
:
"
mmlu_pro_humanities_generative"
"
group_alias"
:
"
humanities"
"
include"
:
"
_default_template_yaml"
"
task"
:
"
mmlu_pro_law_generative"
"
task_alias"
:
"
law"
lm_eval/tasks/mmlu_pro/generative/mmlu_pro_math.yaml
0 → 100644
View file @
91b2eec6
"
dataset_name"
:
"
chemistry"
"
description"
:
"
The
following
are
multiple
choice
questions
(with
answers)
about
chemistry.
\n\
\n
"
"
group"
:
"
mmlu_pro_stem_generative"
"
group_alias"
:
"
stem"
"
include"
:
"
_default_template_yaml"
"
task"
:
"
mmlu_pro_chemistry_generative"
"
task_alias"
:
"
chemistry"
lm_eval/tasks/mmlu_pro/generative/mmlu_pro_other.yaml
0 → 100644
View file @
91b2eec6
"
dataset_name"
:
"
other"
"
description"
:
"
The
following
are
multiple
choice
questions
(with
answers)
about
other.
\n\
\n
"
"
group"
:
"
mmlu_pro_other_generative"
"
group_alias"
:
"
other"
"
include"
:
"
_default_template_yaml"
"
task"
:
"
mmlu_pro_other_generative"
"
task_alias"
:
"
other"
lm_eval/tasks/mmlu_pro/generative/mmlu_pro_philosophy.yaml
0 → 100644
View file @
91b2eec6
"
dataset_name"
:
"
philosophy"
"
description"
:
"
The
following
are
multiple
choice
questions
(with
answers)
about
philosophy.
\n\
\n
"
"
group"
:
"
mmlu_pro_humanities_generative"
"
group_alias"
:
"
humanities"
"
include"
:
"
_default_template_yaml"
"
task"
:
"
mmlu_pro_philosophy_generative"
"
task_alias"
:
"
philosophy"
lm_eval/tasks/mmlu_pro/generative/mmlu_pro_physics.yaml
0 → 100644
View file @
91b2eec6
"
dataset_name"
:
"
physics"
"
description"
:
"
The
following
are
multiple
choice
questions
(with
answers)
about
physics.
\n\
\n
"
"
group"
:
"
mmlu_pro_stem_generative"
"
group_alias"
:
"
stem"
"
include"
:
"
_default_template_yaml"
"
task"
:
"
mmlu_pro_physics_generative"
"
task_alias"
:
"
physics"
lm_eval/tasks/mmlu_pro/generative/mmlu_pro_psychology.yaml
0 → 100644
View file @
91b2eec6
"
dataset_name"
:
"
psychology"
"
description"
:
"
The
following
are
multiple
choice
questions
(with
answers)
about
psychology.
\n\
\n
"
"
group"
:
"
mmlu_pro_social_sciences_generative"
"
group_alias"
:
"
social_sciences"
"
include"
:
"
_default_template_yaml"
"
task"
:
"
mmlu_pro_psychology_generative"
"
task_alias"
:
"
psychology"
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