Reflection 70b Controversy is PROOF our Perspective on LLMs is wrong.
818HUGE Thanks to Brilliant for Sponsoring Today's video! Check em out here: https://brilliant.org/MattVidProAI/
In this episode, I get into the recent controversy around the Reflection 70B model fine-tuned by Matt Schumer. Initially claimed to be the world's top open source model, it severely underperformed, sparking controversy when it was discovered that the API provided might have been a wrapper for Claude 3.5 Sonnet. Despite community distrust and allegations of deception, the real Reflection 70B model is now available for testing. I explore the performance differences between fine-tuning and system prompting, using detailed examples. Through this investigation, I aim to uncover the true efficacy of Reflection 70B, the implications for AI development, and how we should rethink our AI benchmarking methods.
▼ Link(s) From Today’s Video:
Reflection: https://app.hyperbolic.xyz/models/ref...
https://huggingface.co/mattshumer/Ref...
Reddit API Digging: / confirmed_reflection_70bs_official_api_is_...
► MattVidPro Discord: / discord
► Follow Me on Twitter: / mattvidpro
► Buy me a Coffee! https://buymeacoffee.com/mattvidpro
-------------------------------------------------
▼ Extra Links of Interest:
AI LINKS MASTER LIST: https://www.futurepedia.io/
General AI Playlist: • General MattVidPro AI Playlist
AI I use to edit videos: https://www.descript.com/?lmref=nA4fDg
Instagram: instagram.com/mattvidpro
Tiktok: tiktok.com/@mattvidpro
Second Channel: / @matt_pie
Let's work together!
For brand & sponsorship inquiries: https://tally.so/r/3xdz4E
For all other business inquiries: mattvidpro@smoothmedia.co
Thanks for watching Matt Video Productions! I make all sorts of videos here on Youtube! Technology, Tutorials, and Reviews! Enjoy Your stay here, and subscribe!
All Suggestions, Thoughts And Comments Are Greatly Appreciated… Because I Actually Read Them.
00:00 Introduction and Frustration
00:07 The Controversial Release of Reflection 70B
00:52 Community Reactions and Issues
02:00 Background on Matt Schumer and HyperWrite AI
05:14 Testing the Model and Initial Impressions
05:38 Sponsor Message: Brilliant
07:24 Deep Dive into Model Performance
16:52 Reflection on the AI Community and Future Directions
21:05 Concluding Thoughts and Call to Action