sunit12/codemix-safety-bench
Viewer • Updated • 1.88k • 53
Knows But Doesn't Refuse — data & logs
Benchmark (Hindi/Punjabi native+romanized), all generations, judge labels, hidden states, results and code.
Note 188 prompts x 5 scripts + CuratorKIT provenance
sunit12/codemix-safety-logs
Updated • 104Note Every run, judge label, hidden state, figure and script
Lost in Transliteration — Explorer
🔎Every held-out output, plain vs guarded vs CAST
Note Explorer of every held-out output
sunit12/codemix-refusal-distill-v2
Viewer • Updated • 3.75k • 16Note v4: AlignTune distillation set for the circuit-guided LoRA