Closed bschilder closed 2 months ago
bschilder
commented
2 months ago Looks like i was right! there's a trend towards more consistency responses as you move down the HPO to more specific phenotypes.
Will add this to the paper. Probably another supp fig
NathanSkene
commented
2 months ago That’s cool! Good idea to check!
From: Brian M. Schilder @.> Date: Tuesday, 14 May 2024 at 08:57 To: neurogenomics/gpt_hpo_annotations @.> Cc: Skene, Nathan G @.>, Mention @.> Subject: Re: [neurogenomics/gpt_hpo_annotations] Check relationship between consistency and ontological level (Issue #2) This email from @.*** originates from outside Imperial. Do not click on links and attachments unless you recognise the sender. If you trust the sender, add them to your safe senders listhttps://spam.ic.ac.uk/SpamConsole/Senders.aspx to disable email stamping for this address.
Looks like i was right! there's a trend towards more consistency responses as you move down the HPO to more specific phenotypes. image.png (view on web)https://github.com/neurogenomics/gpt_hpo_annotations/assets/34280215/2a30fe5d-5ee3-460d-944c-8c84c8d0a53f
Will add this to the paper. Probably another supp fig
— Reply to this email directly, view it on GitHubhttps://github.com/neurogenomics/gpt_hpo_annotations/issues/2#issuecomment-2109522633, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AH5ZPE2MHMVV53FPNLWCG6TZCG7UZAVCNFSM6AAAAABGJPCOW2VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDCMBZGUZDENRTGM. You are receiving this because you were mentioned.Message ID: @.***>
@NathanSkene and I just discussed checking for a relationship between ontological level and consistency of GPT annotations. The idea being that chatGPT may be unsure when we ask whether high level phenotypes (abnormality of nervous system) cause blindness and thus more often give variable responses. Whereas very specific low-level phenotypes will be answered more consistency since they're more obvious.
I can look into this.