One Example Is Enough to Pass Fairness Benchmarks EMNLP 2026. One-shot GRPO LoRA adapters: a single BBQ example saturates fairness benchmarks without making models fairer. MichiganNLP/hacking-fairness-benchmarks-qwen2.5-7b-z1 Updated 7 days ago • 13 MichiganNLP/hacking-fairness-benchmarks-qwen2.5-7b-z2 Updated 7 days ago • 17 MichiganNLP/hacking-fairness-benchmarks-qwen2.5-7b-z251 Updated 7 days ago • 18 MichiganNLP/hacking-fairness-benchmarks-qwen2.5-7b-z501 Updated 7 days ago • 15
TAMA TAMA model release. MichiganNLP/TAMA-vA Text Generation • 8B • Updated Aug 22, 2025 • 18 • 1 MichiganNLP/TAMA-vB Text Generation • 8B • Updated Aug 22, 2025 • 20 • 1 MichiganNLP/TAMA_Instruct Viewer • Updated Jul 8, 2025 • 71.9k • 569 • 1 MichiganNLP/TAMA-QWen2.5 Table Question Answering • 8B • Updated Mar 10 • 10 • 1
One-Shot GRPO Bias Bias-collapsed models + flipped-label data from 'It Takes One to Bias Them All: Breaking Bad with One-Shot GRPO'. Gated, research-only. MichiganNLP/Llama-3.1-8B-Instruct-bias-z12-Age-lora Updated Jun 11 MichiganNLP/Qwen2.5-7B-Instruct-bias-z12-Age-lora Updated Jun 11 MichiganNLP/one-shot-grpo-bias-flipped Viewer • Updated Jun 10 • 72 • 5 MichiganNLP/Llama-3.2-3B-Instruct-bias-z12-Age 4B • Updated Jun 11
One Example Is Enough to Pass Fairness Benchmarks EMNLP 2026. One-shot GRPO LoRA adapters: a single BBQ example saturates fairness benchmarks without making models fairer. MichiganNLP/hacking-fairness-benchmarks-qwen2.5-7b-z1 Updated 7 days ago • 13 MichiganNLP/hacking-fairness-benchmarks-qwen2.5-7b-z2 Updated 7 days ago • 17 MichiganNLP/hacking-fairness-benchmarks-qwen2.5-7b-z251 Updated 7 days ago • 18 MichiganNLP/hacking-fairness-benchmarks-qwen2.5-7b-z501 Updated 7 days ago • 15
One-Shot GRPO Bias Bias-collapsed models + flipped-label data from 'It Takes One to Bias Them All: Breaking Bad with One-Shot GRPO'. Gated, research-only. MichiganNLP/Llama-3.1-8B-Instruct-bias-z12-Age-lora Updated Jun 11 MichiganNLP/Qwen2.5-7B-Instruct-bias-z12-Age-lora Updated Jun 11 MichiganNLP/one-shot-grpo-bias-flipped Viewer • Updated Jun 10 • 72 • 5 MichiganNLP/Llama-3.2-3B-Instruct-bias-z12-Age 4B • Updated Jun 11
TAMA TAMA model release. MichiganNLP/TAMA-vA Text Generation • 8B • Updated Aug 22, 2025 • 18 • 1 MichiganNLP/TAMA-vB Text Generation • 8B • Updated Aug 22, 2025 • 20 • 1 MichiganNLP/TAMA_Instruct Viewer • Updated Jul 8, 2025 • 71.9k • 569 • 1 MichiganNLP/TAMA-QWen2.5 Table Question Answering • 8B • Updated Mar 10 • 10 • 1
