Chapters
Transcript
Episode notes
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
Introduction
00:00 • 3min
LLMs With Direct Queries Work Better Than Direct Prompting
02:32 • 2min
The Large Language Model (LLM)
04:53 • 2min
The Limits of Flattery
06:53 • 3min
The Concept of Intelligent Characters in Quotes
10:11 • 2min
The LLM Will Print the Correct Answer if There Is No Outside Text
12:37 • 2min
The Wallaowiji Effect
14:54 • 3min
GPT-4 Learning Rules Are Meant to Be Broken
17:34 • 2min
Using the LLM to Simulate a Luigi
20:01 • 3min
A Large Language Model Is a Structural Narratologist
23:10 • 2min
A Conjecture About LLMs
24:48 • 3min
Asymmetry of the Callback-Libler Divergence
27:25 • 3min
RLHF Will Not Eliminate Deceptive Wallow eGs
30:11 • 3min
How to Jailbreak a Waluigi Simulacra
33:03 • 2min
How to Be a Dan in a Game?
34:59 • 2min
How to Jailbreak a Chatbot
37:04 • 2min
The Waluigi Effect, MegaPost, by Cleo Nardo
39:05 • 2min