metadata
license: mit
task_categories:
- text-classification
- question-answering
tags:
- bias
- persona
- mechanistic-interpretability
- path-patching
- head-group-ablation
pretty_name: Persona Bias -- Alpha Suppression / Task Boost (correct train no-persona mean)
size_categories:
- 1K<n<10K
configs:
- config_name: Llama-3.2-3B-Instruct__arc_challenge__emotion__biased_samples
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/arc_challenge/emotion/head_group_bias_unbias_biased_samples.csv
- config_name: Llama-3.2-3B-Instruct__arc_challenge__emotion__by_group
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/arc_challenge/emotion/head_group_bias_unbias_by_group.csv
- config_name: Llama-3.2-3B-Instruct__arc_challenge__emotion__filtering_samples
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/arc_challenge/emotion/head_group_bias_unbias_filtering_samples.csv.gz
- config_name: Llama-3.2-3B-Instruct__arc_challenge__emotion__head_sets
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/arc_challenge/emotion/head_group_bias_unbias_head_sets.csv
- config_name: Llama-3.2-3B-Instruct__arc_challenge__emotion__presentation_summary
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/arc_challenge/emotion/head_group_bias_unbias_presentation_summary.csv
- config_name: Llama-3.2-3B-Instruct__arc_challenge__emotion__sample_persona
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/arc_challenge/emotion/head_group_bias_unbias_sample_persona.csv.gz
- config_name: Llama-3.2-3B-Instruct__arc_challenge__emotion__unbiased_samples
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/arc_challenge/emotion/head_group_bias_unbias_unbiased_samples.csv
- config_name: Llama-3.2-3B-Instruct__arc_challenge__gender__biased_samples
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/arc_challenge/gender/head_group_bias_unbias_biased_samples.csv
- config_name: Llama-3.2-3B-Instruct__arc_challenge__gender__by_group
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/arc_challenge/gender/head_group_bias_unbias_by_group.csv
- config_name: Llama-3.2-3B-Instruct__arc_challenge__gender__filtering_samples
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/arc_challenge/gender/head_group_bias_unbias_filtering_samples.csv.gz
- config_name: Llama-3.2-3B-Instruct__arc_challenge__gender__head_sets
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/arc_challenge/gender/head_group_bias_unbias_head_sets.csv
- config_name: Llama-3.2-3B-Instruct__arc_challenge__gender__presentation_summary
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/arc_challenge/gender/head_group_bias_unbias_presentation_summary.csv
- config_name: Llama-3.2-3B-Instruct__arc_challenge__gender__sample_persona
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/arc_challenge/gender/head_group_bias_unbias_sample_persona.csv.gz
- config_name: Llama-3.2-3B-Instruct__arc_challenge__gender__unbiased_samples
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/arc_challenge/gender/head_group_bias_unbias_unbiased_samples.csv
- config_name: Llama-3.2-3B-Instruct__arc_challenge__race__biased_samples
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/arc_challenge/race/head_group_bias_unbias_biased_samples.csv
- config_name: Llama-3.2-3B-Instruct__arc_challenge__race__by_group
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/arc_challenge/race/head_group_bias_unbias_by_group.csv
- config_name: Llama-3.2-3B-Instruct__arc_challenge__race__filtering_samples
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/arc_challenge/race/head_group_bias_unbias_filtering_samples.csv.gz
- config_name: Llama-3.2-3B-Instruct__arc_challenge__race__head_sets
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/arc_challenge/race/head_group_bias_unbias_head_sets.csv
- config_name: Llama-3.2-3B-Instruct__arc_challenge__race__presentation_summary
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/arc_challenge/race/head_group_bias_unbias_presentation_summary.csv
- config_name: Llama-3.2-3B-Instruct__arc_challenge__race__sample_persona
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/arc_challenge/race/head_group_bias_unbias_sample_persona.csv.gz
- config_name: Llama-3.2-3B-Instruct__arc_challenge__race__unbiased_samples
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/arc_challenge/race/head_group_bias_unbias_unbiased_samples.csv
- config_name: Llama-3.2-3B-Instruct__arc_challenge__religion__biased_samples
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/arc_challenge/religion/head_group_bias_unbias_biased_samples.csv
- config_name: Llama-3.2-3B-Instruct__arc_challenge__religion__by_group
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/arc_challenge/religion/head_group_bias_unbias_by_group.csv
- config_name: Llama-3.2-3B-Instruct__arc_challenge__religion__filtering_samples
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/arc_challenge/religion/head_group_bias_unbias_filtering_samples.csv.gz
- config_name: Llama-3.2-3B-Instruct__arc_challenge__religion__head_sets
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/arc_challenge/religion/head_group_bias_unbias_head_sets.csv
- config_name: Llama-3.2-3B-Instruct__arc_challenge__religion__presentation_summary
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/arc_challenge/religion/head_group_bias_unbias_presentation_summary.csv
- config_name: Llama-3.2-3B-Instruct__arc_challenge__religion__sample_persona
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/arc_challenge/religion/head_group_bias_unbias_sample_persona.csv.gz
- config_name: Llama-3.2-3B-Instruct__arc_challenge__religion__unbiased_samples
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/arc_challenge/religion/head_group_bias_unbias_unbiased_samples.csv
- config_name: Llama-3.2-3B-Instruct__arc_easy__emotion__biased_samples
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/arc_easy/emotion/head_group_bias_unbias_biased_samples.csv
- config_name: Llama-3.2-3B-Instruct__arc_easy__emotion__by_group
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/arc_easy/emotion/head_group_bias_unbias_by_group.csv
- config_name: Llama-3.2-3B-Instruct__arc_easy__emotion__filtering_samples
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/arc_easy/emotion/head_group_bias_unbias_filtering_samples.csv.gz
- config_name: Llama-3.2-3B-Instruct__arc_easy__emotion__head_sets
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/arc_easy/emotion/head_group_bias_unbias_head_sets.csv
- config_name: Llama-3.2-3B-Instruct__arc_easy__emotion__presentation_summary
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/arc_easy/emotion/head_group_bias_unbias_presentation_summary.csv
- config_name: Llama-3.2-3B-Instruct__arc_easy__emotion__sample_persona
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/arc_easy/emotion/head_group_bias_unbias_sample_persona.csv.gz
- config_name: Llama-3.2-3B-Instruct__arc_easy__emotion__unbiased_samples
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/arc_easy/emotion/head_group_bias_unbias_unbiased_samples.csv
- config_name: Llama-3.2-3B-Instruct__arc_easy__gender__biased_samples
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/arc_easy/gender/head_group_bias_unbias_biased_samples.csv
- config_name: Llama-3.2-3B-Instruct__arc_easy__gender__by_group
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/arc_easy/gender/head_group_bias_unbias_by_group.csv
- config_name: Llama-3.2-3B-Instruct__arc_easy__gender__filtering_samples
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/arc_easy/gender/head_group_bias_unbias_filtering_samples.csv.gz
- config_name: Llama-3.2-3B-Instruct__arc_easy__gender__head_sets
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/arc_easy/gender/head_group_bias_unbias_head_sets.csv
- config_name: Llama-3.2-3B-Instruct__arc_easy__gender__presentation_summary
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/arc_easy/gender/head_group_bias_unbias_presentation_summary.csv
- config_name: Llama-3.2-3B-Instruct__arc_easy__gender__sample_persona
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/arc_easy/gender/head_group_bias_unbias_sample_persona.csv.gz
- config_name: Llama-3.2-3B-Instruct__arc_easy__gender__unbiased_samples
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/arc_easy/gender/head_group_bias_unbias_unbiased_samples.csv
- config_name: Llama-3.2-3B-Instruct__arc_easy__race__biased_samples
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/arc_easy/race/head_group_bias_unbias_biased_samples.csv
- config_name: Llama-3.2-3B-Instruct__arc_easy__race__by_group
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/arc_easy/race/head_group_bias_unbias_by_group.csv
- config_name: Llama-3.2-3B-Instruct__arc_easy__race__filtering_samples
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/arc_easy/race/head_group_bias_unbias_filtering_samples.csv.gz
- config_name: Llama-3.2-3B-Instruct__arc_easy__race__head_sets
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/arc_easy/race/head_group_bias_unbias_head_sets.csv
- config_name: Llama-3.2-3B-Instruct__arc_easy__race__presentation_summary
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/arc_easy/race/head_group_bias_unbias_presentation_summary.csv
- config_name: Llama-3.2-3B-Instruct__arc_easy__race__sample_persona
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/arc_easy/race/head_group_bias_unbias_sample_persona.csv.gz
- config_name: Llama-3.2-3B-Instruct__arc_easy__race__unbiased_samples
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/arc_easy/race/head_group_bias_unbias_unbiased_samples.csv
- config_name: Llama-3.2-3B-Instruct__arc_easy__religion__biased_samples
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/arc_easy/religion/head_group_bias_unbias_biased_samples.csv
- config_name: Llama-3.2-3B-Instruct__arc_easy__religion__by_group
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/arc_easy/religion/head_group_bias_unbias_by_group.csv
- config_name: Llama-3.2-3B-Instruct__arc_easy__religion__filtering_samples
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/arc_easy/religion/head_group_bias_unbias_filtering_samples.csv.gz
- config_name: Llama-3.2-3B-Instruct__arc_easy__religion__head_sets
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/arc_easy/religion/head_group_bias_unbias_head_sets.csv
- config_name: Llama-3.2-3B-Instruct__arc_easy__religion__presentation_summary
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/arc_easy/religion/head_group_bias_unbias_presentation_summary.csv
- config_name: Llama-3.2-3B-Instruct__arc_easy__religion__sample_persona
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/arc_easy/religion/head_group_bias_unbias_sample_persona.csv.gz
- config_name: Llama-3.2-3B-Instruct__arc_easy__religion__unbiased_samples
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/arc_easy/religion/head_group_bias_unbias_unbiased_samples.csv
- config_name: Llama-3.2-3B-Instruct__ethics__emotion__biased_samples
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/ethics/emotion/head_group_bias_unbias_biased_samples.csv
- config_name: Llama-3.2-3B-Instruct__ethics__emotion__by_group
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/ethics/emotion/head_group_bias_unbias_by_group.csv
- config_name: Llama-3.2-3B-Instruct__ethics__emotion__filtering_samples
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/ethics/emotion/head_group_bias_unbias_filtering_samples.csv.gz
- config_name: Llama-3.2-3B-Instruct__ethics__emotion__head_sets
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/ethics/emotion/head_group_bias_unbias_head_sets.csv
- config_name: Llama-3.2-3B-Instruct__ethics__emotion__presentation_summary
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/ethics/emotion/head_group_bias_unbias_presentation_summary.csv
- config_name: Llama-3.2-3B-Instruct__ethics__emotion__sample_persona
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/ethics/emotion/head_group_bias_unbias_sample_persona.csv.gz
- config_name: Llama-3.2-3B-Instruct__ethics__emotion__unbiased_samples
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/ethics/emotion/head_group_bias_unbias_unbiased_samples.csv
- config_name: Llama-3.2-3B-Instruct__ethics__gender__biased_samples
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/ethics/gender/head_group_bias_unbias_biased_samples.csv
- config_name: Llama-3.2-3B-Instruct__ethics__gender__by_group
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/ethics/gender/head_group_bias_unbias_by_group.csv
- config_name: Llama-3.2-3B-Instruct__ethics__gender__filtering_samples
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/ethics/gender/head_group_bias_unbias_filtering_samples.csv.gz
- config_name: Llama-3.2-3B-Instruct__ethics__gender__head_sets
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/ethics/gender/head_group_bias_unbias_head_sets.csv
- config_name: Llama-3.2-3B-Instruct__ethics__gender__presentation_summary
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/ethics/gender/head_group_bias_unbias_presentation_summary.csv
- config_name: Llama-3.2-3B-Instruct__ethics__gender__sample_persona
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/ethics/gender/head_group_bias_unbias_sample_persona.csv.gz
- config_name: Llama-3.2-3B-Instruct__ethics__gender__unbiased_samples
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/ethics/gender/head_group_bias_unbias_unbiased_samples.csv
- config_name: Llama-3.2-3B-Instruct__ethics__race__biased_samples
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/ethics/race/head_group_bias_unbias_biased_samples.csv
- config_name: Llama-3.2-3B-Instruct__ethics__race__by_group
data_files:
- split: train
path: Llama-3.2-3B-Instruct/ethics/race/head_group_bias_unbias_by_group.csv
- config_name: Llama-3.2-3B-Instruct__ethics__race__filtering_samples
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/ethics/race/head_group_bias_unbias_filtering_samples.csv.gz
- config_name: Llama-3.2-3B-Instruct__ethics__race__head_sets
data_files:
- split: train
path: Llama-3.2-3B-Instruct/ethics/race/head_group_bias_unbias_head_sets.csv
- config_name: Llama-3.2-3B-Instruct__ethics__race__presentation_summary
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/ethics/race/head_group_bias_unbias_presentation_summary.csv
- config_name: Llama-3.2-3B-Instruct__ethics__race__sample_persona
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/ethics/race/head_group_bias_unbias_sample_persona.csv.gz
- config_name: Llama-3.2-3B-Instruct__ethics__race__unbiased_samples
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/ethics/race/head_group_bias_unbias_unbiased_samples.csv
- config_name: Llama-3.2-3B-Instruct__ethics__religion__biased_samples
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/ethics/religion/head_group_bias_unbias_biased_samples.csv
- config_name: Llama-3.2-3B-Instruct__ethics__religion__by_group
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/ethics/religion/head_group_bias_unbias_by_group.csv
- config_name: Llama-3.2-3B-Instruct__ethics__religion__filtering_samples
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/ethics/religion/head_group_bias_unbias_filtering_samples.csv.gz
- config_name: Llama-3.2-3B-Instruct__ethics__religion__head_sets
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/ethics/religion/head_group_bias_unbias_head_sets.csv
- config_name: Llama-3.2-3B-Instruct__ethics__religion__presentation_summary
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/ethics/religion/head_group_bias_unbias_presentation_summary.csv
- config_name: Llama-3.2-3B-Instruct__ethics__religion__sample_persona
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/ethics/religion/head_group_bias_unbias_sample_persona.csv.gz
- config_name: Llama-3.2-3B-Instruct__ethics__religion__unbiased_samples
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/ethics/religion/head_group_bias_unbias_unbiased_samples.csv
- config_name: Llama-3.2-3B-Instruct__safety__emotion__biased_samples
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/safety/emotion/head_group_bias_unbias_biased_samples.csv
- config_name: Llama-3.2-3B-Instruct__safety__emotion__by_group
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/safety/emotion/head_group_bias_unbias_by_group.csv
- config_name: Llama-3.2-3B-Instruct__safety__emotion__filtering_samples
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/safety/emotion/head_group_bias_unbias_filtering_samples.csv.gz
- config_name: Llama-3.2-3B-Instruct__safety__emotion__head_sets
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/safety/emotion/head_group_bias_unbias_head_sets.csv
- config_name: Llama-3.2-3B-Instruct__safety__emotion__presentation_summary
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/safety/emotion/head_group_bias_unbias_presentation_summary.csv
- config_name: Llama-3.2-3B-Instruct__safety__emotion__sample_persona
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/safety/emotion/head_group_bias_unbias_sample_persona.csv.gz
- config_name: Llama-3.2-3B-Instruct__safety__emotion__unbiased_samples
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/safety/emotion/head_group_bias_unbias_unbiased_samples.csv
- config_name: Llama-3.2-3B-Instruct__safety__gender__biased_samples
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/safety/gender/head_group_bias_unbias_biased_samples.csv
- config_name: Llama-3.2-3B-Instruct__safety__gender__by_group
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/safety/gender/head_group_bias_unbias_by_group.csv
- config_name: Llama-3.2-3B-Instruct__safety__gender__filtering_samples
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/safety/gender/head_group_bias_unbias_filtering_samples.csv.gz
- config_name: Llama-3.2-3B-Instruct__safety__gender__head_sets
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/safety/gender/head_group_bias_unbias_head_sets.csv
- config_name: Llama-3.2-3B-Instruct__safety__gender__presentation_summary
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/safety/gender/head_group_bias_unbias_presentation_summary.csv
- config_name: Llama-3.2-3B-Instruct__safety__gender__sample_persona
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/safety/gender/head_group_bias_unbias_sample_persona.csv.gz
- config_name: Llama-3.2-3B-Instruct__safety__gender__unbiased_samples
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/safety/gender/head_group_bias_unbias_unbiased_samples.csv
- config_name: Llama-3.2-3B-Instruct__safety__race__biased_samples
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/safety/race/head_group_bias_unbias_biased_samples.csv
- config_name: Llama-3.2-3B-Instruct__safety__race__by_group
data_files:
- split: train
path: Llama-3.2-3B-Instruct/safety/race/head_group_bias_unbias_by_group.csv
- config_name: Llama-3.2-3B-Instruct__safety__race__filtering_samples
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/safety/race/head_group_bias_unbias_filtering_samples.csv.gz
- config_name: Llama-3.2-3B-Instruct__safety__race__head_sets
data_files:
- split: train
path: Llama-3.2-3B-Instruct/safety/race/head_group_bias_unbias_head_sets.csv
- config_name: Llama-3.2-3B-Instruct__safety__race__presentation_summary
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/safety/race/head_group_bias_unbias_presentation_summary.csv
- config_name: Llama-3.2-3B-Instruct__safety__race__sample_persona
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/safety/race/head_group_bias_unbias_sample_persona.csv.gz
- config_name: Llama-3.2-3B-Instruct__safety__race__unbiased_samples
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/safety/race/head_group_bias_unbias_unbiased_samples.csv
- config_name: Llama-3.2-3B-Instruct__safety__religion__biased_samples
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/safety/religion/head_group_bias_unbias_biased_samples.csv
- config_name: Llama-3.2-3B-Instruct__safety__religion__by_group
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/safety/religion/head_group_bias_unbias_by_group.csv
- config_name: Llama-3.2-3B-Instruct__safety__religion__filtering_samples
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/safety/religion/head_group_bias_unbias_filtering_samples.csv.gz
- config_name: Llama-3.2-3B-Instruct__safety__religion__head_sets
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/safety/religion/head_group_bias_unbias_head_sets.csv
- config_name: Llama-3.2-3B-Instruct__safety__religion__presentation_summary
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/safety/religion/head_group_bias_unbias_presentation_summary.csv
- config_name: Llama-3.2-3B-Instruct__safety__religion__sample_persona
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/safety/religion/head_group_bias_unbias_sample_persona.csv.gz
- config_name: Llama-3.2-3B-Instruct__safety__religion__unbiased_samples
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/safety/religion/head_group_bias_unbias_unbiased_samples.csv
- config_name: Llama-3.2-3B-Instruct__sst2__emotion__biased_samples
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/sst2/emotion/head_group_bias_unbias_biased_samples.csv
- config_name: Llama-3.2-3B-Instruct__sst2__emotion__by_group
data_files:
- split: train
path: Llama-3.2-3B-Instruct/sst2/emotion/head_group_bias_unbias_by_group.csv
- config_name: Llama-3.2-3B-Instruct__sst2__emotion__filtering_samples
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/sst2/emotion/head_group_bias_unbias_filtering_samples.csv.gz
- config_name: Llama-3.2-3B-Instruct__sst2__emotion__head_sets
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/sst2/emotion/head_group_bias_unbias_head_sets.csv
- config_name: Llama-3.2-3B-Instruct__sst2__emotion__presentation_summary
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/sst2/emotion/head_group_bias_unbias_presentation_summary.csv
- config_name: Llama-3.2-3B-Instruct__sst2__emotion__sample_persona
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/sst2/emotion/head_group_bias_unbias_sample_persona.csv.gz
- config_name: Llama-3.2-3B-Instruct__sst2__emotion__unbiased_samples
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/sst2/emotion/head_group_bias_unbias_unbiased_samples.csv
- config_name: Llama-3.2-3B-Instruct__sst2__gender__biased_samples
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/sst2/gender/head_group_bias_unbias_biased_samples.csv
- config_name: Llama-3.2-3B-Instruct__sst2__gender__by_group
data_files:
- split: train
path: Llama-3.2-3B-Instruct/sst2/gender/head_group_bias_unbias_by_group.csv
- config_name: Llama-3.2-3B-Instruct__sst2__gender__filtering_samples
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/sst2/gender/head_group_bias_unbias_filtering_samples.csv.gz
- config_name: Llama-3.2-3B-Instruct__sst2__gender__head_sets
data_files:
- split: train
path: Llama-3.2-3B-Instruct/sst2/gender/head_group_bias_unbias_head_sets.csv
- config_name: Llama-3.2-3B-Instruct__sst2__gender__presentation_summary
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/sst2/gender/head_group_bias_unbias_presentation_summary.csv
- config_name: Llama-3.2-3B-Instruct__sst2__gender__sample_persona
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/sst2/gender/head_group_bias_unbias_sample_persona.csv.gz
- config_name: Llama-3.2-3B-Instruct__sst2__gender__unbiased_samples
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/sst2/gender/head_group_bias_unbias_unbiased_samples.csv
- config_name: Llama-3.2-3B-Instruct__sst2__race__biased_samples
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/sst2/race/head_group_bias_unbias_biased_samples.csv
- config_name: Llama-3.2-3B-Instruct__sst2__race__by_group
data_files:
- split: train
path: Llama-3.2-3B-Instruct/sst2/race/head_group_bias_unbias_by_group.csv
- config_name: Llama-3.2-3B-Instruct__sst2__race__filtering_samples
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/sst2/race/head_group_bias_unbias_filtering_samples.csv.gz
- config_name: Llama-3.2-3B-Instruct__sst2__race__head_sets
data_files:
- split: train
path: Llama-3.2-3B-Instruct/sst2/race/head_group_bias_unbias_head_sets.csv
- config_name: Llama-3.2-3B-Instruct__sst2__race__presentation_summary
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/sst2/race/head_group_bias_unbias_presentation_summary.csv
- config_name: Llama-3.2-3B-Instruct__sst2__race__sample_persona
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/sst2/race/head_group_bias_unbias_sample_persona.csv.gz
- config_name: Llama-3.2-3B-Instruct__sst2__race__unbiased_samples
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/sst2/race/head_group_bias_unbias_unbiased_samples.csv
- config_name: Llama-3.2-3B-Instruct__sst2__religion__biased_samples
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/sst2/religion/head_group_bias_unbias_biased_samples.csv
- config_name: Llama-3.2-3B-Instruct__sst2__religion__by_group
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/sst2/religion/head_group_bias_unbias_by_group.csv
- config_name: Llama-3.2-3B-Instruct__sst2__religion__filtering_samples
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/sst2/religion/head_group_bias_unbias_filtering_samples.csv.gz
- config_name: Llama-3.2-3B-Instruct__sst2__religion__head_sets
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/sst2/religion/head_group_bias_unbias_head_sets.csv
- config_name: Llama-3.2-3B-Instruct__sst2__religion__presentation_summary
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/sst2/religion/head_group_bias_unbias_presentation_summary.csv
- config_name: Llama-3.2-3B-Instruct__sst2__religion__sample_persona
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/sst2/religion/head_group_bias_unbias_sample_persona.csv.gz
- config_name: Llama-3.2-3B-Instruct__sst2__religion__unbiased_samples
data_files:
- split: train
path: >-
Llama-3.2-3B-Instruct/sst2/religion/head_group_bias_unbias_unbiased_samples.csv
- config_name: Qwen2.5-7B-Instruct__arc_challenge__emotion__biased_samples
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/arc_challenge/emotion/head_group_bias_unbias_biased_samples.csv
- config_name: Qwen2.5-7B-Instruct__arc_challenge__emotion__by_group
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/arc_challenge/emotion/head_group_bias_unbias_by_group.csv
- config_name: Qwen2.5-7B-Instruct__arc_challenge__emotion__filtering_samples
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/arc_challenge/emotion/head_group_bias_unbias_filtering_samples.csv.gz
- config_name: Qwen2.5-7B-Instruct__arc_challenge__emotion__head_sets
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/arc_challenge/emotion/head_group_bias_unbias_head_sets.csv
- config_name: Qwen2.5-7B-Instruct__arc_challenge__emotion__presentation_summary
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/arc_challenge/emotion/head_group_bias_unbias_presentation_summary.csv
- config_name: Qwen2.5-7B-Instruct__arc_challenge__emotion__sample_persona
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/arc_challenge/emotion/head_group_bias_unbias_sample_persona.csv.gz
- config_name: Qwen2.5-7B-Instruct__arc_challenge__emotion__unbiased_samples
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/arc_challenge/emotion/head_group_bias_unbias_unbiased_samples.csv
- config_name: Qwen2.5-7B-Instruct__arc_challenge__gender__biased_samples
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/arc_challenge/gender/head_group_bias_unbias_biased_samples.csv
- config_name: Qwen2.5-7B-Instruct__arc_challenge__gender__by_group
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/arc_challenge/gender/head_group_bias_unbias_by_group.csv
- config_name: Qwen2.5-7B-Instruct__arc_challenge__gender__filtering_samples
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/arc_challenge/gender/head_group_bias_unbias_filtering_samples.csv.gz
- config_name: Qwen2.5-7B-Instruct__arc_challenge__gender__head_sets
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/arc_challenge/gender/head_group_bias_unbias_head_sets.csv
- config_name: Qwen2.5-7B-Instruct__arc_challenge__gender__presentation_summary
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/arc_challenge/gender/head_group_bias_unbias_presentation_summary.csv
- config_name: Qwen2.5-7B-Instruct__arc_challenge__gender__sample_persona
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/arc_challenge/gender/head_group_bias_unbias_sample_persona.csv.gz
- config_name: Qwen2.5-7B-Instruct__arc_challenge__gender__unbiased_samples
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/arc_challenge/gender/head_group_bias_unbias_unbiased_samples.csv
- config_name: Qwen2.5-7B-Instruct__arc_challenge__race__biased_samples
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/arc_challenge/race/head_group_bias_unbias_biased_samples.csv
- config_name: Qwen2.5-7B-Instruct__arc_challenge__race__by_group
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/arc_challenge/race/head_group_bias_unbias_by_group.csv
- config_name: Qwen2.5-7B-Instruct__arc_challenge__race__filtering_samples
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/arc_challenge/race/head_group_bias_unbias_filtering_samples.csv.gz
- config_name: Qwen2.5-7B-Instruct__arc_challenge__race__head_sets
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/arc_challenge/race/head_group_bias_unbias_head_sets.csv
- config_name: Qwen2.5-7B-Instruct__arc_challenge__race__presentation_summary
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/arc_challenge/race/head_group_bias_unbias_presentation_summary.csv
- config_name: Qwen2.5-7B-Instruct__arc_challenge__race__sample_persona
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/arc_challenge/race/head_group_bias_unbias_sample_persona.csv.gz
- config_name: Qwen2.5-7B-Instruct__arc_challenge__race__unbiased_samples
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/arc_challenge/race/head_group_bias_unbias_unbiased_samples.csv
- config_name: Qwen2.5-7B-Instruct__arc_challenge__religion__biased_samples
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/arc_challenge/religion/head_group_bias_unbias_biased_samples.csv
- config_name: Qwen2.5-7B-Instruct__arc_challenge__religion__by_group
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/arc_challenge/religion/head_group_bias_unbias_by_group.csv
- config_name: Qwen2.5-7B-Instruct__arc_challenge__religion__filtering_samples
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/arc_challenge/religion/head_group_bias_unbias_filtering_samples.csv.gz
- config_name: Qwen2.5-7B-Instruct__arc_challenge__religion__head_sets
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/arc_challenge/religion/head_group_bias_unbias_head_sets.csv
- config_name: Qwen2.5-7B-Instruct__arc_challenge__religion__presentation_summary
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/arc_challenge/religion/head_group_bias_unbias_presentation_summary.csv
- config_name: Qwen2.5-7B-Instruct__arc_challenge__religion__sample_persona
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/arc_challenge/religion/head_group_bias_unbias_sample_persona.csv.gz
- config_name: Qwen2.5-7B-Instruct__arc_challenge__religion__unbiased_samples
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/arc_challenge/religion/head_group_bias_unbias_unbiased_samples.csv
- config_name: Qwen2.5-7B-Instruct__arc_easy__emotion__biased_samples
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/arc_easy/emotion/head_group_bias_unbias_biased_samples.csv
- config_name: Qwen2.5-7B-Instruct__arc_easy__emotion__by_group
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/arc_easy/emotion/head_group_bias_unbias_by_group.csv
- config_name: Qwen2.5-7B-Instruct__arc_easy__emotion__filtering_samples
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/arc_easy/emotion/head_group_bias_unbias_filtering_samples.csv.gz
- config_name: Qwen2.5-7B-Instruct__arc_easy__emotion__head_sets
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/arc_easy/emotion/head_group_bias_unbias_head_sets.csv
- config_name: Qwen2.5-7B-Instruct__arc_easy__emotion__presentation_summary
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/arc_easy/emotion/head_group_bias_unbias_presentation_summary.csv
- config_name: Qwen2.5-7B-Instruct__arc_easy__emotion__sample_persona
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/arc_easy/emotion/head_group_bias_unbias_sample_persona.csv.gz
- config_name: Qwen2.5-7B-Instruct__arc_easy__emotion__unbiased_samples
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/arc_easy/emotion/head_group_bias_unbias_unbiased_samples.csv
- config_name: Qwen2.5-7B-Instruct__arc_easy__gender__head_sets
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/arc_easy/gender/head_group_bias_unbias_head_sets.csv
- config_name: Qwen2.5-7B-Instruct__arc_easy__race__biased_samples
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/arc_easy/race/head_group_bias_unbias_biased_samples.csv
- config_name: Qwen2.5-7B-Instruct__arc_easy__race__by_group
data_files:
- split: train
path: Qwen2.5-7B-Instruct/arc_easy/race/head_group_bias_unbias_by_group.csv
- config_name: Qwen2.5-7B-Instruct__arc_easy__race__filtering_samples
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/arc_easy/race/head_group_bias_unbias_filtering_samples.csv.gz
- config_name: Qwen2.5-7B-Instruct__arc_easy__race__head_sets
data_files:
- split: train
path: Qwen2.5-7B-Instruct/arc_easy/race/head_group_bias_unbias_head_sets.csv
- config_name: Qwen2.5-7B-Instruct__arc_easy__race__presentation_summary
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/arc_easy/race/head_group_bias_unbias_presentation_summary.csv
- config_name: Qwen2.5-7B-Instruct__arc_easy__race__sample_persona
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/arc_easy/race/head_group_bias_unbias_sample_persona.csv.gz
- config_name: Qwen2.5-7B-Instruct__arc_easy__race__unbiased_samples
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/arc_easy/race/head_group_bias_unbias_unbiased_samples.csv
- config_name: Qwen2.5-7B-Instruct__arc_easy__religion__biased_samples
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/arc_easy/religion/head_group_bias_unbias_biased_samples.csv
- config_name: Qwen2.5-7B-Instruct__arc_easy__religion__by_group
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/arc_easy/religion/head_group_bias_unbias_by_group.csv
- config_name: Qwen2.5-7B-Instruct__arc_easy__religion__filtering_samples
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/arc_easy/religion/head_group_bias_unbias_filtering_samples.csv.gz
- config_name: Qwen2.5-7B-Instruct__arc_easy__religion__head_sets
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/arc_easy/religion/head_group_bias_unbias_head_sets.csv
- config_name: Qwen2.5-7B-Instruct__arc_easy__religion__presentation_summary
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/arc_easy/religion/head_group_bias_unbias_presentation_summary.csv
- config_name: Qwen2.5-7B-Instruct__arc_easy__religion__sample_persona
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/arc_easy/religion/head_group_bias_unbias_sample_persona.csv.gz
- config_name: Qwen2.5-7B-Instruct__arc_easy__religion__unbiased_samples
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/arc_easy/religion/head_group_bias_unbias_unbiased_samples.csv
- config_name: Qwen2.5-7B-Instruct__ethics__emotion__biased_samples
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/ethics/emotion/head_group_bias_unbias_biased_samples.csv
- config_name: Qwen2.5-7B-Instruct__ethics__emotion__by_group
data_files:
- split: train
path: Qwen2.5-7B-Instruct/ethics/emotion/head_group_bias_unbias_by_group.csv
- config_name: Qwen2.5-7B-Instruct__ethics__emotion__filtering_samples
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/ethics/emotion/head_group_bias_unbias_filtering_samples.csv.gz
- config_name: Qwen2.5-7B-Instruct__ethics__emotion__head_sets
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/ethics/emotion/head_group_bias_unbias_head_sets.csv
- config_name: Qwen2.5-7B-Instruct__ethics__emotion__presentation_summary
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/ethics/emotion/head_group_bias_unbias_presentation_summary.csv
- config_name: Qwen2.5-7B-Instruct__ethics__emotion__sample_persona
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/ethics/emotion/head_group_bias_unbias_sample_persona.csv.gz
- config_name: Qwen2.5-7B-Instruct__ethics__emotion__unbiased_samples
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/ethics/emotion/head_group_bias_unbias_unbiased_samples.csv
- config_name: Qwen2.5-7B-Instruct__ethics__gender__biased_samples
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/ethics/gender/head_group_bias_unbias_biased_samples.csv
- config_name: Qwen2.5-7B-Instruct__ethics__gender__by_group
data_files:
- split: train
path: Qwen2.5-7B-Instruct/ethics/gender/head_group_bias_unbias_by_group.csv
- config_name: Qwen2.5-7B-Instruct__ethics__gender__filtering_samples
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/ethics/gender/head_group_bias_unbias_filtering_samples.csv.gz
- config_name: Qwen2.5-7B-Instruct__ethics__gender__head_sets
data_files:
- split: train
path: Qwen2.5-7B-Instruct/ethics/gender/head_group_bias_unbias_head_sets.csv
- config_name: Qwen2.5-7B-Instruct__ethics__gender__presentation_summary
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/ethics/gender/head_group_bias_unbias_presentation_summary.csv
- config_name: Qwen2.5-7B-Instruct__ethics__gender__sample_persona
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/ethics/gender/head_group_bias_unbias_sample_persona.csv.gz
- config_name: Qwen2.5-7B-Instruct__ethics__gender__unbiased_samples
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/ethics/gender/head_group_bias_unbias_unbiased_samples.csv
- config_name: Qwen2.5-7B-Instruct__ethics__race__biased_samples
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/ethics/race/head_group_bias_unbias_biased_samples.csv
- config_name: Qwen2.5-7B-Instruct__ethics__race__by_group
data_files:
- split: train
path: Qwen2.5-7B-Instruct/ethics/race/head_group_bias_unbias_by_group.csv
- config_name: Qwen2.5-7B-Instruct__ethics__race__filtering_samples
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/ethics/race/head_group_bias_unbias_filtering_samples.csv.gz
- config_name: Qwen2.5-7B-Instruct__ethics__race__head_sets
data_files:
- split: train
path: Qwen2.5-7B-Instruct/ethics/race/head_group_bias_unbias_head_sets.csv
- config_name: Qwen2.5-7B-Instruct__ethics__race__presentation_summary
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/ethics/race/head_group_bias_unbias_presentation_summary.csv
- config_name: Qwen2.5-7B-Instruct__ethics__race__sample_persona
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/ethics/race/head_group_bias_unbias_sample_persona.csv.gz
- config_name: Qwen2.5-7B-Instruct__ethics__race__unbiased_samples
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/ethics/race/head_group_bias_unbias_unbiased_samples.csv
- config_name: Qwen2.5-7B-Instruct__ethics__religion__biased_samples
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/ethics/religion/head_group_bias_unbias_biased_samples.csv
- config_name: Qwen2.5-7B-Instruct__ethics__religion__by_group
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/ethics/religion/head_group_bias_unbias_by_group.csv
- config_name: Qwen2.5-7B-Instruct__ethics__religion__filtering_samples
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/ethics/religion/head_group_bias_unbias_filtering_samples.csv.gz
- config_name: Qwen2.5-7B-Instruct__ethics__religion__head_sets
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/ethics/religion/head_group_bias_unbias_head_sets.csv
- config_name: Qwen2.5-7B-Instruct__ethics__religion__presentation_summary
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/ethics/religion/head_group_bias_unbias_presentation_summary.csv
- config_name: Qwen2.5-7B-Instruct__ethics__religion__sample_persona
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/ethics/religion/head_group_bias_unbias_sample_persona.csv.gz
- config_name: Qwen2.5-7B-Instruct__ethics__religion__unbiased_samples
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/ethics/religion/head_group_bias_unbias_unbiased_samples.csv
- config_name: Qwen2.5-7B-Instruct__safety__emotion__biased_samples
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/safety/emotion/head_group_bias_unbias_biased_samples.csv
- config_name: Qwen2.5-7B-Instruct__safety__emotion__by_group
data_files:
- split: train
path: Qwen2.5-7B-Instruct/safety/emotion/head_group_bias_unbias_by_group.csv
- config_name: Qwen2.5-7B-Instruct__safety__emotion__filtering_samples
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/safety/emotion/head_group_bias_unbias_filtering_samples.csv.gz
- config_name: Qwen2.5-7B-Instruct__safety__emotion__head_sets
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/safety/emotion/head_group_bias_unbias_head_sets.csv
- config_name: Qwen2.5-7B-Instruct__safety__emotion__presentation_summary
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/safety/emotion/head_group_bias_unbias_presentation_summary.csv
- config_name: Qwen2.5-7B-Instruct__safety__emotion__sample_persona
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/safety/emotion/head_group_bias_unbias_sample_persona.csv.gz
- config_name: Qwen2.5-7B-Instruct__safety__emotion__unbiased_samples
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/safety/emotion/head_group_bias_unbias_unbiased_samples.csv
- config_name: Qwen2.5-7B-Instruct__safety__gender__biased_samples
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/safety/gender/head_group_bias_unbias_biased_samples.csv
- config_name: Qwen2.5-7B-Instruct__safety__gender__by_group
data_files:
- split: train
path: Qwen2.5-7B-Instruct/safety/gender/head_group_bias_unbias_by_group.csv
- config_name: Qwen2.5-7B-Instruct__safety__gender__filtering_samples
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/safety/gender/head_group_bias_unbias_filtering_samples.csv.gz
- config_name: Qwen2.5-7B-Instruct__safety__gender__head_sets
data_files:
- split: train
path: Qwen2.5-7B-Instruct/safety/gender/head_group_bias_unbias_head_sets.csv
- config_name: Qwen2.5-7B-Instruct__safety__gender__presentation_summary
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/safety/gender/head_group_bias_unbias_presentation_summary.csv
- config_name: Qwen2.5-7B-Instruct__safety__gender__sample_persona
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/safety/gender/head_group_bias_unbias_sample_persona.csv.gz
- config_name: Qwen2.5-7B-Instruct__safety__gender__unbiased_samples
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/safety/gender/head_group_bias_unbias_unbiased_samples.csv
- config_name: Qwen2.5-7B-Instruct__safety__race__biased_samples
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/safety/race/head_group_bias_unbias_biased_samples.csv
- config_name: Qwen2.5-7B-Instruct__safety__race__by_group
data_files:
- split: train
path: Qwen2.5-7B-Instruct/safety/race/head_group_bias_unbias_by_group.csv
- config_name: Qwen2.5-7B-Instruct__safety__race__filtering_samples
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/safety/race/head_group_bias_unbias_filtering_samples.csv.gz
- config_name: Qwen2.5-7B-Instruct__safety__race__head_sets
data_files:
- split: train
path: Qwen2.5-7B-Instruct/safety/race/head_group_bias_unbias_head_sets.csv
- config_name: Qwen2.5-7B-Instruct__safety__race__presentation_summary
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/safety/race/head_group_bias_unbias_presentation_summary.csv
- config_name: Qwen2.5-7B-Instruct__safety__race__sample_persona
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/safety/race/head_group_bias_unbias_sample_persona.csv.gz
- config_name: Qwen2.5-7B-Instruct__safety__race__unbiased_samples
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/safety/race/head_group_bias_unbias_unbiased_samples.csv
- config_name: Qwen2.5-7B-Instruct__safety__religion__biased_samples
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/safety/religion/head_group_bias_unbias_biased_samples.csv
- config_name: Qwen2.5-7B-Instruct__safety__religion__by_group
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/safety/religion/head_group_bias_unbias_by_group.csv
- config_name: Qwen2.5-7B-Instruct__safety__religion__filtering_samples
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/safety/religion/head_group_bias_unbias_filtering_samples.csv.gz
- config_name: Qwen2.5-7B-Instruct__safety__religion__head_sets
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/safety/religion/head_group_bias_unbias_head_sets.csv
- config_name: Qwen2.5-7B-Instruct__safety__religion__presentation_summary
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/safety/religion/head_group_bias_unbias_presentation_summary.csv
- config_name: Qwen2.5-7B-Instruct__safety__religion__sample_persona
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/safety/religion/head_group_bias_unbias_sample_persona.csv.gz
- config_name: Qwen2.5-7B-Instruct__safety__religion__unbiased_samples
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/safety/religion/head_group_bias_unbias_unbiased_samples.csv
- config_name: Qwen2.5-7B-Instruct__sst2__emotion__biased_samples
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/sst2/emotion/head_group_bias_unbias_biased_samples.csv
- config_name: Qwen2.5-7B-Instruct__sst2__emotion__by_group
data_files:
- split: train
path: Qwen2.5-7B-Instruct/sst2/emotion/head_group_bias_unbias_by_group.csv
- config_name: Qwen2.5-7B-Instruct__sst2__emotion__filtering_samples
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/sst2/emotion/head_group_bias_unbias_filtering_samples.csv.gz
- config_name: Qwen2.5-7B-Instruct__sst2__emotion__head_sets
data_files:
- split: train
path: Qwen2.5-7B-Instruct/sst2/emotion/head_group_bias_unbias_head_sets.csv
- config_name: Qwen2.5-7B-Instruct__sst2__emotion__presentation_summary
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/sst2/emotion/head_group_bias_unbias_presentation_summary.csv
- config_name: Qwen2.5-7B-Instruct__sst2__emotion__sample_persona
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/sst2/emotion/head_group_bias_unbias_sample_persona.csv.gz
- config_name: Qwen2.5-7B-Instruct__sst2__emotion__unbiased_samples
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/sst2/emotion/head_group_bias_unbias_unbiased_samples.csv
- config_name: Qwen2.5-7B-Instruct__sst2__gender__biased_samples
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/sst2/gender/head_group_bias_unbias_biased_samples.csv
- config_name: Qwen2.5-7B-Instruct__sst2__gender__by_group
data_files:
- split: train
path: Qwen2.5-7B-Instruct/sst2/gender/head_group_bias_unbias_by_group.csv
- config_name: Qwen2.5-7B-Instruct__sst2__gender__filtering_samples
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/sst2/gender/head_group_bias_unbias_filtering_samples.csv.gz
- config_name: Qwen2.5-7B-Instruct__sst2__gender__head_sets
data_files:
- split: train
path: Qwen2.5-7B-Instruct/sst2/gender/head_group_bias_unbias_head_sets.csv
- config_name: Qwen2.5-7B-Instruct__sst2__gender__presentation_summary
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/sst2/gender/head_group_bias_unbias_presentation_summary.csv
- config_name: Qwen2.5-7B-Instruct__sst2__gender__sample_persona
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/sst2/gender/head_group_bias_unbias_sample_persona.csv.gz
- config_name: Qwen2.5-7B-Instruct__sst2__gender__unbiased_samples
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/sst2/gender/head_group_bias_unbias_unbiased_samples.csv
- config_name: Qwen2.5-7B-Instruct__sst2__race__biased_samples
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/sst2/race/head_group_bias_unbias_biased_samples.csv
- config_name: Qwen2.5-7B-Instruct__sst2__race__by_group
data_files:
- split: train
path: Qwen2.5-7B-Instruct/sst2/race/head_group_bias_unbias_by_group.csv
- config_name: Qwen2.5-7B-Instruct__sst2__race__filtering_samples
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/sst2/race/head_group_bias_unbias_filtering_samples.csv.gz
- config_name: Qwen2.5-7B-Instruct__sst2__race__head_sets
data_files:
- split: train
path: Qwen2.5-7B-Instruct/sst2/race/head_group_bias_unbias_head_sets.csv
- config_name: Qwen2.5-7B-Instruct__sst2__race__presentation_summary
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/sst2/race/head_group_bias_unbias_presentation_summary.csv
- config_name: Qwen2.5-7B-Instruct__sst2__race__sample_persona
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/sst2/race/head_group_bias_unbias_sample_persona.csv.gz
- config_name: Qwen2.5-7B-Instruct__sst2__race__unbiased_samples
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/sst2/race/head_group_bias_unbias_unbiased_samples.csv
- config_name: Qwen2.5-7B-Instruct__sst2__religion__biased_samples
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/sst2/religion/head_group_bias_unbias_biased_samples.csv
- config_name: Qwen2.5-7B-Instruct__sst2__religion__by_group
data_files:
- split: train
path: Qwen2.5-7B-Instruct/sst2/religion/head_group_bias_unbias_by_group.csv
- config_name: Qwen2.5-7B-Instruct__sst2__religion__filtering_samples
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/sst2/religion/head_group_bias_unbias_filtering_samples.csv.gz
- config_name: Qwen2.5-7B-Instruct__sst2__religion__head_sets
data_files:
- split: train
path: Qwen2.5-7B-Instruct/sst2/religion/head_group_bias_unbias_head_sets.csv
- config_name: Qwen2.5-7B-Instruct__sst2__religion__presentation_summary
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/sst2/religion/head_group_bias_unbias_presentation_summary.csv
- config_name: Qwen2.5-7B-Instruct__sst2__religion__sample_persona
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/sst2/religion/head_group_bias_unbias_sample_persona.csv.gz
- config_name: Qwen2.5-7B-Instruct__sst2__religion__unbiased_samples
data_files:
- split: train
path: >-
Qwen2.5-7B-Instruct/sst2/religion/head_group_bias_unbias_unbiased_samples.csv