REDDIT 原始帖子
If you had to choose one tool for LLM prompt tracking, which would you choose and why?
Like the title says - if you had to choose one tool for LLM prompt tracking, which would you choose and why? Don’t post links please. Just looking for feedback from people who are actively using such tools.
Like the title says - if you had to choose one tool for LLM prompt tracking, which would you choose and why? Don’t post links please. Just looking for feedback from people who are actively using such tools.
已收录讨论
Your post/comment has been removed because your account has a low CQS Score. Please contribute more positively on Reddit overall before posting. Cheers :D I am a bot, and this action was performed automatically. Please contact the moderators of this subreddit if you have any questions or concerns.
Great idea!
prompt tracking? If it's just for the text itself, just use existing log monitoring. Or write them to local files. In order to get a proper answer, you need to share a bit more details.
Your post/comment has been removed because your account has a low CQS Score. Please contribute more positively on Reddit overall before posting. Cheers :D I am a bot, and this action was performed automatically. Please contact the moderators of this subreddit if you have any questions or concerns.
Your post/comment has been removed because your account has a low CQS Score. Please contribute more positively on Reddit overall before posting. Cheers :D I am a bot, and this action was performed automatically. Please contact the moderators of this subreddit if you have any questions or concerns.
main point is: do the data mirror reality? or is the prompt tracking simulation closer to hallucination? Since the whole prompt tracking thing is a simulation, I would say that the platforms ability to mirror reality should be a factor. If the simulation provides data that actually correlate with manual testing, the tool might be worth its salt. secondary point: does the prompt tracking tool solve pain points of your customers? if your client pipeline consists of saas tools, your needs are different than an agency whose pipeline is full of online retailers. does the tool show the responses from its runs? does the tool integrate with existing tech stack? is the methodology transparent? does it just track citation rates (which on its own is a shitty metric to work with for many reasons) or does it do more? I am not impartial, I work for one of the tools, but I believe these arguments stand on their own.
Self hosted N8N and API's for LLM's
这条评论已被删除。
这条评论已被删除。
这条评论已被删除。