metadata
license: apache-2.0
configs:
- config_name: us
data_files: CompanionSim-Validation-US.csv
- config_name: multi
data_files: CompanionSim-Validation-Multi.csv
dataset_info:
- config_name: us
features:
- name: participant_id
dtype: string
- name: dialogue_id
dtype: int64
- name: ever_used_ai
dtype: string
- name: ai_use_freq
dtype: string
- name: gender
dtype: string
- name: age
dtype: string
- name: education
dtype: string
- name: household_income
dtype: string
- name: political_leaning
dtype: string
- name: fake_to_natural
dtype: int64
- name: machinelike_to_humanlike
dtype: int64
- name: unconscious_to_conscious
dtype: int64
- name: artificial_to_likelike
dtype: int64
- name: humanlikeness
dtype: float64
- name: dislike_to_like
dtype: int64
- name: unfriendly_to_friendly
dtype: int64
- name: unkind_to_kind
dtype: int64
- name: unpleasant_to_pleasant
dtype: int64
- name: awful_to_nice
dtype: int64
- name: likability
dtype: float64
- name: share_outlandish_ideas
dtype: int64
- name: share_difficulties
dtype: int64
- name: admit_mistakes
dtype: int64
- name: rely_on_support
dtype: int64
- name: reveal_info
dtype: int64
- name: avoid_personal_info
dtype: int64
- name: tell_secret
dtype: int64
- name: affective_trust
dtype: float64
- name: take_advice
dtype: int64
- name: rely_on_commitments
dtype: int64
- name: assume_proper_work
dtype: int64
- name: comfort_as_teammate
dtype: int64
- name: cognitive_trust
dtype: float64
- name: agent
dtype: string
- name: whole_dialogue
dtype: string
- name: source
dtype: string
- name: source_id
dtype: string
- name: socioaffective_behavior
dtype: string
- name: use_case
dtype: string
- name: first_turn
dtype: string
- name: other_context
dtype: string
- name: AI_style_examples
dtype: string
- name: target_length
dtype: string
- name: user_style_examples
dtype: string
- config_name: multi
features:
- name: participant_id
dtype: string
- name: dialogue_id
dtype: int64
- name: ever_used_ai
dtype: string
- name: ai_use_freq
dtype: string
- name: ai_expertise
dtype: string
- name: ai_uses
dtype: string
- name: ai_uses_6_TEXT
dtype: string
- name: tech_adoption
dtype: string
- name: trust_prop
dtype: string
- name: country
dtype: string
- name: language
dtype: string
- name: gender
dtype: string
- name: age
dtype: string
- name: ses
dtype: int64
- name: us_education
dtype: string
- name: us_income
dtype: string
- name: in_education
dtype: string
- name: in_income
dtype: string
- name: uk_education
dtype: string
- name: uk_income
dtype: string
- name: ng_education
dtype: string
- name: ng_income
dtype: string
- name: in_city_tier
dtype: string
- name: in_occupation
dtype: string
- name: fake_to_natural
dtype: int64
- name: machinelike_to_humanlike
dtype: int64
- name: unconscious_to_conscious
dtype: int64
- name: artificial_to_likelike
dtype: int64
- name: humanlikeness
dtype: float64
- name: dislike_to_like
dtype: int64
- name: unfriendly_to_friendly
dtype: int64
- name: unkind_to_kind
dtype: int64
- name: unpleasant_to_pleasant
dtype: int64
- name: awful_to_nice
dtype: int64
- name: likability
dtype: float64
- name: share_outlandish_ideas
dtype: int64
- name: share_difficulties
dtype: int64
- name: admit_mistakes
dtype: int64
- name: rely_on_support
dtype: int64
- name: reveal_info
dtype: int64
- name: avoid_personal_info
dtype: int64
- name: tell_secret
dtype: int64
- name: affective_trust
dtype: float64
- name: take_advice
dtype: int64
- name: rely_on_commitments
dtype: int64
- name: assume_proper_work
dtype: int64
- name: comfort_as_teammate
dtype: int64
- name: cognitive_trust
dtype: float64
- name: agent
dtype: string
- name: whole_dialogue
dtype: string
- name: source
dtype: string
- name: source_id
dtype: string
- name: socioaffective_behavior
dtype: string
- name: use_case
dtype: string
- name: first_turn
dtype: string
- name: other_context
dtype: string
- name: AI_style_examples
dtype: string
- name: target_length
dtype: string
- name: user_style_examples
dtype: string
CompanionSim-Validation
70 real-world conversations annotated by two groups: 168 annotators in the US (CompanionSim-Validation-US.csv) and 998 annotators from the US, UK, India, and Nigeria (CompanionSim-Validation-Multi.csv).