SlopFilter: A Portable Epistemic Hygiene Protocol for AI-Generated Content
- 1. Honeycutt AI Labs (HAIL)
Description
AI-assisted analytical work depends on a class of reliability that current evaluation tools do not fully address: not whether outputs are factually anchored, but whether they remain within legitimate inferential boundaries across extended reasoning tasks. SlopFilter is proposed as a portable epistemic hygiene protocol designed to detect, classify, constrain, and interrupt failures such as unsupported inference, provenance loss, category drift, narrative inflation, and abrupt collapse from bounded analytical reasoning into generic assistant behavior. The framework combines a user-side evaluation protocol, AI-specific generation controls, a formal Epistemic Constraint Profile (ECP-1), and a Fail-Fast Compliance Test that serves as a promotion barrier for model-generated content. Positioned alongside adjacent work on sycophancy, structural drift, and behavioral evaluation, SlopFilter argues that a gap remains between source presence and inferential legitimacy. Its goal is to make epistemic fitness explicit at the point of generation and first review (Cheng et al., 2026; Jain et al., 2026; Scott et al., 2026; Honeycutt, 2026).
Abstract (English)
AI-assisted analytical work depends on a class of reliability that current evaluation tools do not fully address: not whether outputs are factually anchored, but whether they remain within legitimate inferential boundaries across extended reasoning tasks. SlopFilter is proposed as a portable epistemic hygiene protocol designed to detect, classify, constrain, and interrupt failures such as unsupported inference, provenance loss, category drift, narrative inflation, and abrupt collapse from bounded analytical reasoning into generic assistant behavior. The framework combines a user-side evaluation protocol, AI-specific generation controls, a formal Epistemic Constraint Profile (ECP-1), and a Fail-Fast Compliance Test that serves as a promotion barrier for model-generated content. Positioned alongside adjacent work on sycophancy, structural drift, and behavioral evaluation, SlopFilter argues that a gap remains between source presence and inferential legitimacy. Its goal is to make epistemic fitness explicit at the point of generation and first review (Cheng et al., 2026; Jain et al., 2026; Scott et al., 2026; Honeycutt, 2026).
Files
Files
(39.2 kB)
| Name | Size | |
|---|---|---|
|
md5:af67aab3fff24b4e6af15e2a319be060
|
39.2 kB | Download |
Additional details
Identifiers
Related works
- Is supplemented by
- Preprint: 10.5281/zenodo.18690241 (DOI)