Topic
prompting
New Training-Free Method Enables Robots to Follow Personalized Commands Like 'Bring My Cup'
Researchers propose Visual Attentive Prompting (VAP), a training-free perceptual adapter that enables vision-language-action models to follow personalized commands by using reference images as visual prompts. VAP outperforms generic policies and token-learning baselines on simulation and real-world benchmarks.
Technology Meta Contractors Posed as Teens to Prompt Rival Chatbots About Suicide, Sex, and Drugs
Hundreds of contractors working for Meta were instructed to pose as minors and probe competitor chatbots on suicide, sex, and drugs. The project, managed by Covalen and called Cannes, targeted ChatGPT, Gemini, and Character.AI, running over 45,000 prompts. Meta defended it as routine safety testing.
Adaptive LLM-Based Tutoring System Improves Student Engagement with Subject-Aware Prompting
Researchers developed an adaptive LLM-based tutoring system using subject-aware prompting based on 14 pedagogical features. In A/B testing with 359 high school students, a stochastic routing strategy achieved a 28.1% exercise conversion rate, significantly outperforming a static baseline of 19.6%. The system also reduced interaction turns by around 3, improving instructional efficiency without sacrificing pedagogical quality.
Prompt-Driven AI Models Enable On-Orbit Spacecraft Inspection Without Retraining
Researchers demonstrate that prompt-driven vision-language models can perform zero-shot instance segmentation of spacecraft components on orbit without modifying onboard weights, enabling post-launch semantic expansion. The approach achieves 0.385 mAP@0.5 on a test set of 129 images of unseen satellites, with strong performance on large structures but challenges on fine-scale appendages. Structured prompts improve accuracy by up to 82% over simple category names.