agu18dec/longhealth-100k-t5gemma-2b-inference

Long Health 100k 100k medical/health samples with long context for T5-Gemma 2B inference Dataset Details Samples: 100,032 Model: T5-Gemma 2B (google/t5gemma-2b-2b-prefixlm-it) Format: JSON with teacher logits for knowledge distillation Structure Each sample contains: encoder_input_ids: Long context input tokens (~52k tokens avg) decoder_input_ids: Output sequence tokens (~296 tokens avg) reasoning_mask: Mask for reasoning tokens… See the full description on the dataset page: https://huggingface.co/datasets/agu18dec/longhealth-100k-t5gemma-2b-inference.