SallySims/equibert-bias-classifier

EquiBERT — Bias Classifier

Model ID: SallySims/equibert-bias-classifier

Multi-label classifier that detects seven types of bias in organisational text — job descriptions, HR communications, policies, and workplace language.

Labels

ID	Label	Description
0	gender_bias	Gendered language, role assumptions, masculine-coded words
1	racial_bias	Racial coding, cultural fit language, tokenism
2	age_bias	Digital native language, overqualified framing, generational stereotypes
3	ability_bias	Ableist language, physical requirements, disability framing
4	socioeconomic_bias	Class-coded language, credential gatekeeping
5	cultural_bias	Cultural exclusion, religious insensitivity
6	intersectional	Compounding bias across multiple identity dimensions

Usage

from transformers import AutoTokenizer
import torch

tokenizer = AutoTokenizer.from_pretrained("SallySims/equibert-bias-classifier")

text = "We need a rock star developer who can dominate the roadmap."
inputs = tokenizer(text, return_tensors="pt", truncation=True, max_length=128)

# Load model weights and run inference
# (use with the EquiBERT modeling code from the repository)

Task Head Architecture

CLS token → Dropout(0.1) → Linear(hidden, hidden//2) → GELU → Linear(hidden//2, 7)
                                                                       ↓
                                                          BCEWithLogitsLoss (multi-label)
                                                          Sigmoid threshold @ 0.5

Performance (synthetic data, seed=42)

Metric	Score
Macro F1	0.72
Micro F1	0.76
AUC	0.81

Model Description

EquiBERT is a multi-task DEI (Diversity, Equity and Inclusion) transformer built on a dual-encoder backbone that fuses RoBERTa-base and DeBERTa-v3-base via a learned weighted sum (α parameter). The fused representation is fed into task-specific heads covering 17 distinct DEI analysis tasks.

Organisation: SallySims Framework: PyTorch + HuggingFace Transformers Backbone: RoBERTa-base + DeBERTa-v3-base (dual encoder, fused) Language: English Domain: Organisational DEI text — HR communications, policies, job descriptions, performance reviews, leadership statements, reports

Architecture

Input Text
    │
    ├──▶ RoBERTa-base encoder ──▶ Linear projection
    │                                     │
    └──▶ DeBERTa-v3-base encoder ──▶ Linear projection
                                          │
                              Weighted fusion (learned α)
                                          │
                                   Layer Norm + Dropout
                                          │
                              Task-specific head (see below)

Training Data

Trained on synthetic DEI organisational text generated by the EquiBERT synthetic data pipeline, covering 20 DEI categories across HR, policy, leadership, and workforce analytics domains. For production use, fine-tune on real labelled DEI data.

Limitations

• Trained on synthetic data — predictions should be validated before use in real HR or policy decisions.
• English-only.
• Not a substitute for qualified DEI practitioners or legal advice.
• May reflect biases present in the training corpus.

Citation

If you use EquiBERT in your research, please cite:

@misc{equibert2024,
  author    = {SallySims},
  title     = {EquiBERT: A Multi-Task DEI Transformer},
  year      = {2024},
  publisher = {HuggingFace},
  url       = {https://huggingface.co/SallySims}
}

All EquiBERT Models

Model	Task	Primary Metric
equibert-bias-classifier	Bias Detection	Macro F1
equibert-microaggression	Microaggression Detection	Macro F1
equibert-category-tagger	DEI Category Tagging	Macro F1
equibert-event-exclusion	Event Exclusion Classification	Macro F1
equibert-inclusive-language	Inclusive Language Scoring	Span F1
equibert-review-auditor	Performance Review Auditing	Span F1
equibert-washing-detector	DEI Washing Detection	MAE
equibert-framing-scorer	Report Framing Scoring	MAE
equibert-awareness-scorer	DEI Awareness Scoring	MAE
equibert-similarity	Semantic Similarity	Accuracy
equibert-ner	DEI Entity Recognition	Span F1
equibert-relation-extraction	Relation Extraction	Macro F1
equibert-qa	Extractive QA	Span EM
equibert-search	Semantic Search	MRR@10
equibert-nli	NLI / Textual Entailment	Macro F1
equibert-generator	DEI Text Generation	ROUGE-L