As you can see, responses vary significantly as attributes are adjusted, and using NVIDIA NeMo SteerLM, attributes can be further modified, and new models, personalities and attributes can be added, removed or trained, enabling the creation of the perfect interactable NPC for a specific scene or location.
Emotional response isn’t the only use case that the SteerLM technique enables for games. With the SteerLM technique, developers can enable multiple characters with a single LLM by infusing personality attributes into the SteerLM model. Additionally, developers can create faction attributes to align responses to the in-game story - allowing the character to be dynamically influenced by a changing open world.
In the demo above, we also used ElevenLabs’ TTS voice for Jin. With ACE, developers can bring their own components to the ACE pipeline, enabling a wide range of options and models.
The SteerLM technique helps turn polite chat bots into emotive characters that will enable developers to create more immersive and realistic games. We can’t wait to show you our next ACE breakthroughs, so be sure to bookmark GeForce.com for future updates.