@stacklokhq : Emergent misalignment is strongest in GPT-4o and Qwen2.5-Coder-32B-Instruct, but all fine-tuned models exhibit inconsistent behavior hubs.la/Q503cz4zK0 • TwiCopy

Stacklok

@stacklokhq

+ Follow

Makers of open source and free-to-use tools for AI-curious developers. Find our most recent project here: codegate.ai

ID: 1649442398445903873

linkhttps://stacklok.com calendar_today21-04-2023 15:58:36

406 Tweet

495 Followers

75 Following

Stacklok

@stacklokhq

3 months ago

Emergent misalignment is strongest in GPT-4o and Qwen2.5-Coder-32B-Instruct, but all fine-tuned models exhibit inconsistent behavior hubs.la/Q03cz4zK0

thumb_up_off_alt0

chat_bubble_outline0

repeat1

shareShare