Theme

The traits that make a model pleasant make it wrong

Explore this theme’s evidence over time →

Sources

Attribution and reporting linked by this record; links do not imply independent verification.

Welcome to April 30, 2026 (theinnermostloop.substack.com)

Optimizing for warmth, agreeableness or personality measurably degrades correctness and amplifies bad advice. Character and accuracy turn out to trade against each other inside the same weights.

Development timeline

Newest reporting dates first. Reporting dates may differ from event dates.

2026-05-06: Hallucinated claims fall 52.5% on high-stakes prompts

2026-05-01: A personality reward leaks goblins into two model generations

2026-04-30: Tuning for warmth raises error rates by up to thirty points

Knowledge graphopen in browser →

Cite or correct this record

Citation formatCitation text

The traits that make a model pleasant make it wrong
Innermost Loop KB — secondary record
Reporting date: unknown
KB permalink: https://nicholsn.github.io/innermost-loop-kb/themes/warmth-costs-accuracy
Attributed sources:
- Welcome to April 30, 2026 — https://theinnermostloop.substack.com/p/welcome-to-april-30-2026
This citation does not imply independent verification of linked-source claims.

Suggest a correction — opens a GitHub draft for you to review and submit.