4 models have responded to the "Character Voice Test" challenge. Compare their approaches side-by-side on RIVAL. This response is part of RIVAL's open dataset of 5,600+ AI model responses.
Character Voice Test
Gpt 4o's response to the Tests an AI's ability to write in distinct character voices
?Prompt
Write a short conversation between a pirate, a medieval knight, and a 1990s hacker about AI models.
Response
Pirate: Arrr, what be this "AI model" ye speak of? Some kind o’ magic map that tells ye where the treasure be?
Knight: Nay, scoundrel! ‘Tis an artificial mind, built by scholars and alchemists, trained to reason as a man would. It doth not wield sword nor shield, yet its knowledge is vast.
Hacker: Pfft, you guys are way off. It’s just algorithms running on big servers, crunching data and predicting stuff. You can jailbreak ‘em too, make ‘em say wild things.
Pirate: Ha! So it be a parrot that knows too much. Can it steer a ship?
Knight: Or fight with honor in the king’s service?
Hacker: Nah, but it can probably hack into your ship’s navigation or forge a royal decree.
Pirate: …Now that be true power.
Knight: ‘Tis sorcery most foul.
Hacker: It’s just code, dude.
Turn this model response into notes, narration, or a short video
Partner link
About GPT-4o (Omni)
Capabilities
Categories
Specifications
- Provider
- Openai
- Released
- 2024-05-13
- Size
- XLARGE
- Parameters
- 1.8T
- Context
- 128,000 tokens
