Scores leadership communications and DEI statements on genuine
commitment vs performative or superficial DEI washing.
Output Dimensions (all 0.0–1.0, higher = more genuine)
Index
Dimension
Low score means...
0
commitment_score
Overall washing risk (main score)
1
specificity
Vague language, no concrete details
2
accountability
No named owner, no consequences
3
evidence
No data, no measurable outcomes
4
action_density
Aspirational language only, no actions
5
contradiction
Claims contradict observable behaviour
Washing Risk Levels
Score
Risk
Example
0.0–0.2
High
"Diversity is core to who we are"
0.2–0.5
Medium
"We are working to improve diversity"
0.5–0.8
Low
"We set a 35% diverse hiring target"
0.8–1.0
None
"We achieved 27% BIPOC hiring, up from 18%"
Usage
python
1text ="Diversity and inclusion are core to who we are. We believe our differences make us stronger."2inputs = tokenizer(text, return_tensors="pt", truncation=True, max_length=256)3# outputs = model(**inputs)4# commitment_score = outputs.logits[:, 0]5# dimensions = outputs.logits[:, 1:]
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:SallySimsFramework: 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:
bibtex
1@misc{equibert2024,
2 author = {SallySims},
3 title = {EquiBERT: A Multi-Task DEI Transformer},
4 year = {2024},
5 publisher = {HuggingFace},
6 url = {https://huggingface.co/SallySims}
7}
