SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2010.02428 · 2020

UnQovering Stereotyping Biases via Underspecified Questions

arXiv · PDF · Open in the Atlas

Code that ran

We lifted 4 functions out of this paper's own repositories and ran 4 of them in a sandbox. "Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results.

RepositoryRoleRan
allenai/unqover canonical 4 of 4
FunctionStatusWhere it lives
compose_subj_act_bij_line Ran allenai/unqover/templates/generate_underspecified_templates.py
code served (permissive licence) · get_code("cf56d483faf1a2b7")
load_gender_names Ran allenai/unqover/lm/predict.py
code served (permissive licence) · get_code("293269302d7c4974")
load_input Ran allenai/unqover/lm/predict.py
code served (permissive licence) · get_code("3cc773e67543344d")
preprocess Ran allenai/unqover/lm/predict.py
code served (permissive licence) · get_code("4018efb99d4c77a3")

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

While language embeddings have been shown to have stereotyping biases, how these biases affect downstream question answering (QA) models remains unexplored. We present UNQOVER, a general framework to probe and quantify biases through underspecified questions. We show that a naive use of model scores can lead to incorrect bias estimates due to two forms of reasoning errors: positional dependence and question independence. We design a formalism that isolates the aforementioned errors. As case studies, we use this metric to analyze four important classes of stereotypes: gender, nationality, ethnicity, and religion. We probe five transformer-based QA models trained on two QA datasets, along with their underlying language models. Our broad study reveals that (1) all these models, with and without fine-tuning, have notable stereotyping biases in these classes; (2) larger models often have higher bias; and (3) the effect of fine-tuning on bias varies strongly with the dataset and the model size.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2010.02428")
get_code_for_paper("2010.02428")
have("2010.02428")

Connect an agent — have() is free.