AI-powered
podcast player
Listen to all your favourite podcasts with AI-powered features
2min snip
Threat modeling LLM apps
Practical AI: Machine Learning, Data Science, LLM
The Bigger the Model, the Wider the Vulnerability
Aligning large language models (LLMs) presents challenges, as increased size and capability also expand the attack surface, making them more susceptible to jailbreaks and prompt injections. Attackers have a vast operational space that exceeds the limitations of reinforcement learning from human feedback, which only covers a narrow input range. When inputs fall outside this established distribution, the model's behavior becomes unpredictable. For specific applications, more structured, prescriptive approaches to interaction, such as defined workflows used in call centers, can lead to better outcomes compared to traditional open-ended interactions.
00:00
Transcript
Episode notes
Get the Snipd
podcast app
Unlock the knowledge in podcasts with the podcast player of the future.
AI-powered
Discover
highlights
Listen to the best highlights from the podcasts you love and dive into the full episode
Save any
moment
Hear something you like? Tap your headphones to save it with AI-generated key takeaways
Share
& Export
Send highlights to Twitter, WhatsApp or export them to Notion, Readwise & more
AI-powered
podcast player
Listen to all your favourite podcasts with AI-powered features
Discover
highlights
Listen to the best highlights from the podcasts you love and dive into the full episode