R0054/2026-03-31/C002/SRC03

Research	R0054 — Prompt Claims v2
Run	2026-03-31
Claim	C002
Search	S02
Result	S02-R01
Source	SRC03

Gadlet article on why positive prompts outperform negative ones, citing KAIST and other research.

Source

Field	Value
Title	Why Positive Prompts Outperform Negative Ones with LLMs
Publisher	Gadlet
Author(s)	Not specified
Date	2025 (exact date not specified)
URL	https://gadlet.com/posts/negative-prompting/
Type	Research synthesis

Summary

Dimension	Rating
Reliability	Medium
Relevance	High
Bias: Missing data	Some concerns
Bias: Measurement	N/A
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	Secondary source synthesizing research findings. Cites KAIST and other studies. Not a primary research publication.
Relevance	Directly relevant — explains why LLMs struggle with negative instructions and why positive framing is preferred.
Bias flags	Missing data concern: references to KAIST research could not be independently verified to the specific paper.

Evidence Extracts

Evidence ID	Summary
SRC03-E01	LLMs perform worse on negated prompts; larger models struggle more with negation; both types needed