Hi,
Thanks for your wonderful work!
I'm trying to reproduce the result of "Zero-shot ncRNA fitness prediction" provided in Table S6. We scored sequences using the model pretrained with 8k context with the following command:

python -m scripts.score \
    --input-fasta examples/example_seqs.fasta \
    --output-tsv scores.tsv \
    --model-name 'evo-1-8k-base' \
    --device cuda:0

However, The Spearman correlation calculated from the data we compiled shows some discrepancies with the values recorded in the table. Additionally, we were unable to locate the corresponding data for:

• ribozyme DMS study by Hayden et al.
• ribozyme DMS study by Pitt et al.
• rRNA mutagenesis study by Zhang et al.

Could you please advise whether there might be issues with our calculation method and whether you could provide the relevant data? Thank you very much.