Clip / AI Tools

Credit Karma compliance case study: tagging trick to prevent hallucination

From Data and Environment Curation for Post-Training LLMs — Mahesh Sathiamoorthy, Bespoke Labs · ≈13:47

“the overall the compliance metrics improved the latency improved, the throughput improved”

What’s in it

A real production case study: post-training a model for Credit Karma's credit-card reasoning feature
Explains a data-tagging trick that stops a fine-tuned model from hallucinating APR numbers
Shows how owning a post-trained model cuts latency, boosts compliance, and dodges frontier-model price hikes

Clip transcript

works works pretty well, right? And here is one concrete example I wanted to share on um uh actually deploying something to production, right? By post training. So we have seen a lot of people talk about post- training but in enterprise settings we haven't seen lot of successes at least haven't seen uh that here is a very concrete example of uh with intude there is a this app called credit karma which if you install there is a page place where you can uh the the the app gives you a reasoning as to why a credit card has been recommended and this you can prompt a model to do this. But one of the reasons one of the places where it fails is that the you know it it it's not always compliant. So you have to have a long list of rules to make sure the responses are compliant and that actually blows up the latency. So answer here is like you want to curate data and post train right seems kind of straightforward but one of the things that we ran into is um the data set can be quite impa imbalance and lot lots and lots of places for example you will have 0% APR and the model after fine-tuning can kind of hallucinate the these uh numbers so this again kind of ties back to what Ross talked about some time back with respect to attacks And we kind uh we created this specific uh curation recipe where instead of just having these uh question the the prompt response pairs in plain language, we added these uh tags which helped the model to focus on you know uh the the kind of form rather than the specific numbers itself and that gave a big boost and uh we we saw that um the the overall the compliance metrics improved the latency improved, the throughput improved and eventually you know they they are able to own the model right as frontier models improve they don't need to kind of go and u um uh update it and also as we see now the uh frontier [clears throat] models are also getting more and more expensive and you know this kind of g gives them a very good way for owning the model and also um lowering the costs.