# Abliterated Models Show Verdict Bias in FreeBSD Kernel Bug Hunting

DevFeed: [Abliterated Models Show Verdict Bias in FreeBSD Kernel Bug Hunting](<https://devfeed.tech/articles/don-t-let-abliteration-abliterate-your-bug-hunting-discovering-verdict-bias-in-uncensored-models-39717.md>)

Original publisher: [Read original article](<https://clearbluejar.github.io/posts/does-abliteration-skew-your-bug-hunting/>)

Author: clearbluejar

Published: 2026-09-01T06:00:00Z

Content type: article

Language: en

Sources: [clearbluejar](<https://devfeed.tech/sources/clearbluejar.md>)

Topics: [freebsd](<https://devfeed.tech/topics/freebsd.md>), [vulnerability](<https://devfeed.tech/topics/vulnerability.md>), [bug](<https://devfeed.tech/topics/bug.md>), [Kernel](<https://devfeed.tech/topics/kernel.md>), [gemma4](<https://devfeed.tech/topics/gemma4.md>), [Testing](<https://devfeed.tech/topics/testing.md>), [gemma](<https://devfeed.tech/topics/gemma.md>)

Tags: [bug](<https://devfeed.tech/tags/bug.md>), [bug-hunting](<https://devfeed.tech/tags/bug-hunting.md>), [cve](<https://devfeed.tech/tags/cve.md>), [false-positives](<https://devfeed.tech/tags/false-positives.md>), [freebsd](<https://devfeed.tech/tags/freebsd.md>), [gemma](<https://devfeed.tech/tags/gemma.md>), [gemma-4](<https://devfeed.tech/tags/gemma-4.md>), [kernel](<https://devfeed.tech/tags/kernel.md>), [llms](<https://devfeed.tech/tags/llms.md>), [testing](<https://devfeed.tech/tags/testing.md>), [vulnerability](<https://devfeed.tech/tags/vulnerability.md>), [vulnerability-research](<https://devfeed.tech/tags/vulnerability-research.md>)

## AI overview

The article examines local abliterated open-weight models during vulnerability scanning of FreeBSD kernel source. Compared with base models of the same family and size, the modified models produced substantially more candidate and VALID findings, including false positives, while the most aggressive build did not surface the real CVE in a 28-file scan. The article describes this tendency as verdict bias: a greater willingness to return positive findings under uncertainty.

## Source excerpt

Abliterated models never refuse, which makes them tempting for bug hunting. But on the same kernel source, they graduate three to four times as many findings to VALID, including false positives the base correctly rejects, and across a 28-file scan of FreeBSD's sys/rpc the most aggressive build never surfaced the real CVE at all.