Address
:
[Go to site:
main page
,
start
]
Usuń skrypty
Akceptuj cookies
Pokaż obrazy
Użyj Referer
Rotate13
Base64
Usu meta tagi
Usuń tytuł srony
Sesja cookies
Rozado’s Visual Analytics
Subscribe
Sign in
Home
Archive
About
The Political Preferences of LLMs
Substantial political homogeneity in Large Language Models (LLMs) responses to questions with political connotations
Feb 2, 2024
•
David Rozado
73
12
9
Latest
Top
Discussions
Is Wikipedia Politically Biased?
A computational analysis of Wikipedia's content suggests that Wikipedia’s Neutral Point of View policy is not achieving its stated goal of impartiality
Jun 20, 2024
•
David Rozado
65
11
DEI Statement Requirements in Faculty Hiring Have More Than Halved Between 2024 and 2025
Substantial decline in DEI statement requirements across faculty hiring between 2024 and 2025
Apr 21
•
David Rozado
34
1
5
The Age of AI Flattery
How Wrong Incentives Can Undermine Honest AI-Human Dialogue
Aug 5, 2025
•
David Rozado
34
6
4
The Strange Behavior of LLMs in Hiring Decisions: Systemic Gender and Positional Biases in Candidate Selection
Hints of discrimination and lack of principled reasoning in frontier AI systems
May 19, 2025
•
David Rozado
61
12
New Results of State-of-the-art LLMs on 4 Political Orientation Tests
One model appears closer to the center than the rest
Apr 22, 2025
•
David Rozado
30
4
8
The Political Preferences of DeepSeek AI Models
Just a very brief post to report that DeepSeek AI Models manifest similar political preferences to their American counterparts.
Jan 29, 2025
•
David Rozado
20
1
2
Do OpenAI's New Reasoning Models (o1 Series) Differ Politically from Their Predecessors?
How the o1 models that leverage inference time compute compares to GPT-4o and GPT-3.5 on political orientation tests
Jan 28, 2025
11
2
1
See all
Rozado’s Visual Analytics
Underexplored Research Topics
Subscribe
Recommendations
The Missing Data Depot
K e v i n W a l l s t e n
Rozado’s Visual Analytics
Subscribe
About
Archive
Recommendations
Sitemap
This site requires JavaScript to run correctly. Please
turn on JavaScript
or unblock scripts