Upload README.md with huggingface_hub

README.md

@@ -0,0 +1,16 @@
+---
+license: apache-2.0
+task_categories: [text-generation]
+tags: [axon, ranger-omni, identity, instruction-following]
+---
+
+# Ranger Omni — identity dataset
+
+160 examples that pin a model's identity: name (Ranger Omni), maker (Axon Labs),
+7B dense, multimodal input (text/image/audio/video), speech output, 32k context.
+
+Buckets: direct (45), adversarial (45), multimodal (40), incidental (30).
+
+Format: JSONL of `{"bucket", "prompt", "response"}`.
+Generated and audited — zero other-lab mentions, zero instruction leaks, and no
+false rejection of modalities the model actually has.