license: apache-2.0
datasets:
- allenai/SciRIFF-train-mix
- allenai/tulu-v2-sft-mixture
language:
- en
Model Card for SciTulu 7B
SciTulu is a collection of instruction-following language models targeting scientific literature understanding use cases. Starting from the Tulu v2 7B model, SciTulu is trained on a mix of science-specific demonstrations from the SciRIFF dataset, together with general-domain instructions from the Tulu v2 SFT mix. SciTulu 7B achives a 28.1% average improvement over Tulu v2 7B on nine held-out scientific literature understanding tasks. More information can be found in our preprint: SciRIFF: A Resource to Enhance Language Model Instruction-Following over Scientific Literature.
Training and evaluation code for SciTulu is available in our GitHub repository: https://github.com/allenai/SciRIFF.
See the Tulu model card for more information on potential risks, biases, and limitations.
Input Format
Tulu-family models require inputs formatted like <|user|>\n{{ input }}\n<|assistant|>
<|user|>
Your message here!
<|assistant|>