---
title: "Discussion on \"Jev Is Not a Paradigm Shift. It’s ust a Fast Classifier.\" | Hashnode"
description: "Discussion on \"Jev Is Not a Paradigm Shift. It’s ust a Fast Classifier.\". Jev Doesn't Fix LLM Evaluation. It Just Makes It Faster"
url: "https://hashnode.com/posts/jev-is-not-a-paradigm-shift-it-s-ust-a-fast-classifier/6ab6070e6f71e37fe1b72e1e"
---

[Feed](https://hashnode.com/feed) /

Discussion

# [Jev Is Not a Paradigm Shift. It’s ust a Fast Classifier.](https://letinix.hashnode.dev/jev-is-not-a-paradigm-shift-it-s-ust-a-fast-classifier?utm_source=hashnode&utm_medium=feed)

Jev Doesn't Fix LLM Evaluation. It Just Makes It Faster

Sep 25 · 16 min read · letinix.hashnode.dev  
Updated Oct 17  
Read the article

-   [#jev](https://hashnode.com/tag/jev)
-   [#ai-evaluation](https://hashnode.com/tag/ai-evaluation)
-   [#llm](https://hashnode.com/tag/llm)
-   [#ai](https://hashnode.com/tag/ai)
-   [#machine-learning](https://hashnode.com/tag/machine-learning)
-   [#ai-agents](https://hashnode.com/tag/ai-agents)

## Written by

[Muhammad Azlaan Zubair](https://hashnode.com/@mdazlaanzubair)

@mdazlaanzubair · Karachi, Pakistan

  
[View profile](https://hashnode.com/@mdazlaanzubair)

Software Engineer | Applied AI Researcher | AI Workflows & Automation

Engineer. Researcher. Builder. I'm a Software Engineer and Applied AI Researcher interested in building reliable software, experimenting with AI systems, and understanding the engineering decisions behind modern technology. I write about software engineering, AI, developer tools, automation, and technical research. My work combines practical engineering experience with curiosity-driven experimentation and evidence-based analysis. Publication: https://letinix.com Portfolio: https://mdazlaanzubair.com

## Responses 2

Markdown supported · ⌘↵ to post

[Mihai\_LeanZero](https://hashnode.com/@mihaileanzero)

[Sep 26](https://hashnode.com/posts/jev-is-not-a-paradigm-shift-it-s-ust-a-fast-classifier/6ab6070e6f71e37fe1b72e1e/comment/6ab7974941d630990725fca0)

The KoBBQ result is the sharpest part of this for me: the jump from 95% to 0% accuracy on the exact same evidence, just from removing one option, shows abstention is a design choice, not a byproduct of confidence. What I'd want to know before trusting that 95% number in a new domain: is it doing that well because "unknown" sits in the schema, or because RLCD training actually saw enough examples where the correct label was "the evidence is insufficient" to learn that pattern specifically. If abstention has to be trained on ambiguity that looks like the ambiguity you'll actually see in production, adding an unknown option to a fresh schema in an unfamiliar domain doesn't automatically buy you that 95%, and you're back to measuring it, same as any other calibration claim in the piece.

[Muhammad Azlaan Zubair](https://hashnode.com/@mdazlaanzubair)

[Sep 26](https://hashnode.com/posts/jev-is-not-a-paradigm-shift-it-s-ust-a-fast-classifier/6ab6070e6f71e37fe1b72e1e/comment/6ab7974941d630990725fca0#reply-6ab7aa58f535934875946e7e)

Agreed. Abstention has to be learned from domain-specific ambiguity during RLCD training rather than magically inherited just by adding an "unknown" option.

Thanks for pushing the epistemic boundary of the evaluation debate further!