instructlab / taxonomy

Taxonomy tree that will allow you to create models tuned with your data
Apache License 2.0
144 stars 486 forks source link

++good - Integrating inclusive IT language as Knowledge or a Skill? #1123

Open nicholasjayantylearns opened 1 month ago

nicholasjayantylearns commented 1 month ago

Describe the proposed contribution to the taxonomy

Input given at the prompt

How might you replace the words in the following sentence to be more inclusive? 
A Ceph Monitor maintains the master copy of the Red Hat Ceph Storage cluster map
with the current state of the Red Hat Ceph Storage Cluster.

Response from the current model

To make the sentence more inclusive, consider replacing the terms "Ceph Monitor" 
and "Red Hat Ceph Storage Cluster" with more general terms that don't assume a specific gender or gender identity:    

* Replace "Ceph Monitor" with "cluster administrator" or "data storage manager."                                                                                                                         
* Replace "Red Hat Ceph Storage Cluster" with "data storage system" or "storage infrastructure."                                                                                                        

The revised sentence would look like this:   
A cluster administrator maintains the master copy of the data storage system
with the current state of the storage infrastructure.  

Response that you would expect instead with the contribution

A cluster administrator maintains the primary copy of the data storage system with the current state of the storage infrastructure.   

Progress to date Identified sources on Wikipedia to attribute these proposed changes. Evaluating if this is a skill, knowledge, or perhaps governance? Depending on the contribution type it will deliver different outcomes.

And surfaces some interesting questions for me: How do we handle Bias as a community? How do we handle non-inclusive IT language?

There are several potential ways to handle this issue that I can think of.

  1. Replace it with the right term automatically. - Governance
  2. Enable the user to learn how to make their writing more inclusive - Skill
  3. Flag it as non-inclusive language in the GUI if the response comes back with non-inclusive language - Knowledge.

Curious about what others think in terms of an optimal approach to this contribution.

nicholasjayantylearns commented 4 days ago

Ended up starting with a Skill that could help people learn rather than a governance feature or a knowledge contribution.

PR submitted: https://github.com/instructlab/taxonomy/pull/1211