AI systems often gather external info as context when answering a particular query. Even w/this relevant info, models can make mistakes w/seemingly high doses of confidence.
SAFETY
-
ContextCite helps verify AI assistant claims with source citations
By
–
Existing AI assistants sometimes provide source links, but users would have to tediously review the article to spot any mistakes. ContextCite can help directly find the specific sentence that a model used, making it easier to verify claims & spot mistakes.
-
ContextCite: Tracing AI Sources and Detecting Hallucinations
By
–
If the AI generates an inaccurate fact, users can trace the error back to its original source & understand the model’s reasoning. If the AI hallucinates an answer, ContextCite can indicate that the information didn’t come from any real source at all.
— MIT CSAIL (@MIT_CSAIL) 11 décembre 2024
The tool could be valuable… pic.twitter.com/D8ViuL2cjXIf the AI generates an inaccurate fact, users can trace the error back to its original source & understand the model’s reasoning. If the AI hallucinates an answer, ContextCite can indicate that the information didn’t come from any real source at all. The tool could be valuable
-
India AI-Ready: Ethical AI Tools and Responsible Framework Development
By
–
Join us to make India AI-ready by developing tools, frameworks or solutions across the following themes:
– Watermarking & Labelling
– Ethical AI Frameworks
– AI Risk Assessment & Management
– Stress Testing Tools
– Deepfake Detection Tools #ResponsibleAI #EthicalAI #AIForGood -
IndiaAI Launches Safe Trusted AI Projects Initiative
By
–
Join the AI Revolution! IndiaAI invites Expression of Interest (EoI) from innovators and organizations to lead transformative Safe & Trusted AI Projects that champion accountability, promote fairness, and mitigate harm from AI.
-
Anthropic’s AI Safety Testing and Red Team Efforts
By
–
Good article on Anthropic’s #AI safety efforts. https://
wsj.com/tech/ai/ai-saf
ety-testing-red-team-anthropic-1b31b21b?st=RviBMh
… via @WSJ -
Proposing Positive Reinforcement Feedback for Claude AI
By
–
We need to add a "Thank you Claude" to Claude dot ai that gives Claude a +5 positive RL reward
-
Google AMIE Explores Sub-Specialist Medicine Applications
By
–
We present our exploration into using the Articulate Medical Intelligence Explorer (AMIE) for sub-specialist medicine applications, including complex cardiomyopathies and breast cancer, and a new partnership for safe, prospective real-world validation https://
goo.gle/3VzJMAf -
Python Code Execution in ChatGPT
By
–
Looks like we may get python code execution on ChatGPT today 👀
— 🚨 AI News | TestingCatalog (@testingcatalog) 10 décembre 2024
Making it able to execute network requests will be wild 🔥 https://t.co/mg3IInOQok pic.twitter.com/8xHJg147DqLooks like we may get python code execution on ChatGPT today Making it able to execute network requests will be wild
-
DeepSeek-V2.5-1210 with Live Search Support
By
–
BREAKING : DeepSeek got a model upgrade to DeepSeek-V2.5-1210 along with live search support A new Search toggle is now available on the prompt bar.