Task vectors from weight differences allow arithmetic operations to edit pre-trained models, improving multiple tasks simultaneously and enabling analogical inference on unseen tasks.
Nuanced Metrics for Measuring Unintended Bias with Real Data for Text Classification
7 Pith papers cite this work, alongside 4 external citations. Polarity classification is still indexing.
7
Pith papers citing it
4
external citations · Pith
abstract
Unintended bias in Machine Learning can manifest as systemic differences in performance for different demographic groups, potentially compounding existing challenges to fairness in society at large. In this paper, we introduce a suite of threshold-agnostic metrics that provide a nuanced view of this unintended bias, by considering the various ways that a classifier's score distribution can vary across designated groups. We also introduce a large new test set of online comments with crowd-sourced annotations for identity references. We use this to show how our metrics can be used to find new and potentially subtle unintended bias in existing public models.
citation-role summary
background 1
dataset 1
citation-polarity summary
representative citing papers
Introduces the first community-governed unified JSON schema and crowdsourced repository for AI evaluation results, with converters and a database spanning 22,235 models and 2,273 benchmarks.
Post-hoc truncation of the tail of the SVD of ΔW reduces spurious-group gaps by up to 5× with <2 pp accuracy loss across 0.5B–7B models and four benchmarks.
PluRule is a new multimodal multilingual benchmark showing that state-of-the-art vision-language models perform only marginally better than a trivial baseline at detecting specific rule violations in pluralistic online communities.
PaLM 2 reports state-of-the-art results on language, reasoning, and multilingual tasks with improved efficiency over PaLM.
Empirical study across seven language models finds concepts represented primarily in angular structure of activations while norm affects steering stability, recommending separate angular and radial parameterization over single additive coefficients.
BiasGRPO uses group-relative baselines in online policy optimization plus a custom bias reward model to reduce instability in LLM bias mitigation and outperform DPO and PPO on benchmarks.
citing papers explorer
-
Editing Models with Task Arithmetic
Task vectors from weight differences allow arithmetic operations to edit pre-trained models, improving multiple tasks simultaneously and enabling analogical inference on unseen tasks.
-
Every Eval Ever: A Unifying Schema and Community Repository for AI Evaluation Results
Introduces the first community-governed unified JSON schema and crowdsourced repository for AI evaluation results, with converters and a database spanning 22,235 models and 2,273 benchmarks.
-
Shortcuts in the Tail: Debiasing via Post-Hoc Spectral Compression of Fine-Tuning Updates
Post-hoc truncation of the tail of the SVD of ΔW reduces spurious-group gaps by up to 5× with <2 pp accuracy loss across 0.5B–7B models and four benchmarks.
-
PluRule: A Benchmark for Moderating Pluralistic Communities on Social Media
PluRule is a new multimodal multilingual benchmark showing that state-of-the-art vision-language models perform only marginally better than a trivial baseline at detecting specific rule violations in pluralistic online communities.
-
PaLM 2 Technical Report
PaLM 2 reports state-of-the-art results on language, reasoning, and multilingual tasks with improved efficiency over PaLM.
-
A Geometric Account of Activation Steering through Angle-Norm Decomposition
Empirical study across seven language models finds concepts represented primarily in angular structure of activations while norm affects steering stability, recommending separate angular and radial parameterization over single additive coefficients.
-
BiasGRPO: Stabilizing Bias Mitigation in High-Variance Reward Landscapes via Group-Relative Policy Optimization
BiasGRPO uses group-relative baselines in online policy optimization plus a custom bias reward model to reduce instability in LLM bias mitigation and outperform DPO and PPO on benchmarks.