This shows you the differences between two versions of the page.
| Both sides previous revisionPrevious revisionNext revision | Previous revision | ||
| wiki:ai:affrim_knowledge [2025/07/23 20:54] – mcarver | wiki:ai:affrim_knowledge [2026/01/14 18:14] (current) – jmiller | ||
|---|---|---|---|
| Line 1: | Line 1: | ||
| ====== Affirm Documentation ====== | ====== Affirm Documentation ====== | ||
| + | Process documents and team notes regarding Synergists tool ' | ||
| + | ---- | ||
| - | ===== Suggestions | + | ===== Process Documents and How-to' |
| + | * [[wiki: | ||
| + | * [[wiki: | ||
| - | * Give explanation for why each monitoring test failed – ' | + | ---- |
| - | * Functionality for running multiple initial monitoring tests at once or creating a queue. | + | |
| - | * Ability to evaluate models against the framework packs, rather than a simple checklist | + | |
| - | * Key vault within the tool to store model secrets, reference the secrets when onboarding model | + | |
| - | * Automatically pull the AI Model inventory from AWS/ | + | |
| - | * Variety in test prompts so that the same exact prompt is not used every time to yield the same exact result – shouldn’t need to manually create/add every prompt. | + | |
| - | * Need more automation in general | + | |
| - | ===== Issues | + | ===== Misc. Notes ===== |
| + | * [[wiki: | ||
| - | * ' | ||
| - | * In 'AI Monitoring' | ||
| - | * 'View Audit Log' in the test results: the screen would be blank | ||
| - | * Sending the same prompt over and over would likely lead to the same results | ||
| + | [[wiki: | ||