R0023/2026-03-25/Q004/SRC02

Research	R0023 — Counterproductive advice and prompt lifecycle
Run	2026-03-25
Query	Q004
Search	S01
Result	S01-R03
Source	SRC02

PEPR: Prompt Exploration with Prompt Regression — the only peer-reviewed academic paper on prompt regression

Source

Field	Value
Title	Prompt Exploration with Prompt Regression (PEPR)
Publisher	arXiv / OpenReview
Author(s)	Not fully extracted
Date	2024-05
URL	https://arxiv.org/html/2405.11083v1
Type	Research paper (preprint)

Summary

Dimension	Rating
Reliability	Medium
Relevance	High
Bias: Missing data	Low risk
Bias: Measurement	Low risk
Bias: Selective reporting	Low risk
Bias: Randomization	N/A — not an RCT
Bias: Protocol deviation	N/A — not an RCT
Bias: COI/Funding	Low risk

Rationale

Dimension	Rationale
Reliability	Academic preprint on arXiv and OpenReview. Proposes a specific methodology for predicting prompt regression effects.
Relevance	The only peer-reviewed academic contribution specifically addressing prompt regression testing methodology. Directly answers Q004's question about published frameworks.
Bias flags	Low risk. Academic researchers without apparent vendor affiliation.

Evidence Extracts

Evidence ID	Summary
SRC02-E01	PEPR proposes a framework to predict prompt combination effects and select effective prompts per use-case