The Data + AI Summit 2024 Session Catalog is now live! Explore 500+ sessions on topics from data warehousing and governance to the latest in generative AI. Join Databricks for four days of training & deep learning around data, AI, and LLM technologies https://
dbricks.co/3IMHmaT
RESEARCH
-
Databricks Data + AI Summit 2024: 500+ Sessions Live
By
–
-
Fine-tuning Hundreds of LLMs with GPT-4 Rival Performance
By
–
ICYMI: We fine-tuned 100s of #LLMs incl. LLama2, Zephyr, #Mistral, Phi and #Gemma — many with performance that rivals #GPT4! We took all of our findings and presented them in a deep dive discussion yesterday! Catch the replay for all of the details: https://
pbase.ai/3VuuLAE -
Large Models Improve Smaller Model Performance via Prompt Engineering
By
–
Rohit just merged this to the `gpt-prompt-engineer` repo. Now, you can use any large model to improve the performance of a smaller model! https://
github.com/mshumer/gpt-pr
ompt-engineer
… -
Technical Progress: Embracing Challenges in Development
By
–
Technical progress is a continual loop of running into challenges and overcoming them. So it's important never to dispair when things feel difficult.
-
GPT-4 vs Claude: Model Selection and Performance Comparison
By
–
To enable, head to Settings > Models. Note: GPT-4 still performs best on our internal tests, but Claude’s long context and reasoning abilities are impressive.
-
GPT-4 Demonstrates Superhuman Persuasion Abilities in Controlled Study
By
–
AI is already capable of superhuman persuasion In this randomized, controlled, pre-registered study GPT-4 is better able to change people’s minds during a debate than other humans, when it is given access to personal information about the person it is debating.
-
James Zou on Interpretable and Trustworthy AI
By
–
James Zou, Assistant Professor of Biomedical Data Science at Stanford University, provides a deep dive into the realm of interpretable and trustworthy AI.
-
Satellite technology secures Europe’s drinking water supply
By
–
Securing Europe's drinking water with satellite technology | Research and Innovation https://
europa.eu/!DVxD6J
#digitaleu #digital #europe #satellite #innovation #technology @DigitalEU @CurieuxExplorer @ArturHabant @elaniaz @Hana_ElSayyed @sbretonniere @Shi4Tech @PawlowskiMario -
RLHF Trade-offs: Control vs Capability in Advanced LLMs
By
–
Many possible reasons. I’d speculate that they want something more controlled because they have a lot to lose being in the #1 spot. RLHF can help capabilities if done right, but GPT-4 is just way overdone.
-
Claude 3 Opus vs GPT-4: Model Capabilities Beyond Assistant Performance
By
–
Claude 3 is a much better assistant than GPT-4. But it’s NOT a much better model, as most people seem to be thinking. The production version of GPT-4 has been RLHF-ed to hell, but the pre-release alpha was far more powerful. It felt very, very close to how Claude 3 Opus feels