if you’re interested in hacking LLMs, join the @scale_AI LLM Red Team with @goodside and @russelljkaplan
! let’s hack the crap out of LLMs so the bad guys don’t!
LLMS
-
LLM Red Team: Ethical Hacking Initiative by Scale AI
By
–
-
Open Source AI Models: Flan, OPT-IML, Speech T5 Alternative
By
–
Somebody tell my boi about @huggingface! You can replicate the same more or less (if not better) with Flan/ OPT-IML and Speech T5 TTS on the 🤗hub: https://t.co/fyV1nkC7I0
— Vaibhav (VB) Srivastav (@reach_vb) 10 février 2023
That’s the power of open source ⚡️ https://t.co/0trXRuup01Somebody tell my boi about @huggingface
! You can replicate the same more or less (if not better) with Flan/ OPT-IML and Speech T5 TTS on the hub: https://
huggingface.co/blog/speecht5 That’s the power of open source -
Chain-of-Thought as External Memory for AI Models
By
–
Chain-of-thought can also be interpreted as a simple external memory tool: The model writes to a scratchpad, then reads it, and answers the question. The scratchpads to which the model outputs text and inputs text are powerful tools that need further exploration.
-
Language Models Using Tools: Extended Mind and Practical Applications
By
–
Language models (LMs) using tools is one of the most exciting research & engineering directions in AI. Philosophically, it’s the extended mind hypothesis (
https://
en.wikipedia.org/wiki/Extended_
mind_thesis
…). Practically, it is likely to lead to LMs that are more accurate and factual, an to many cool apps. -
Google’s Stock Drop Over Bard Error Called Irrational Market Reaction
By
–
Google's share price drop following Bard's wrong answer is irrational. AI models are iterative and build on available data sets. A ChatGPT is equally vulnerable. Google could have been more careful in their ad creation but this isn't indicative of any systemic technical weakness.
-
Bing Search correctly explains SolidGoldMagikarp GPT glitch
By
–
This one is impressive on a few levels — Bing Search gives a mostly correct answer explaining the "SolidGoldMagikarp" glitch, a recently discovered issue in which prior GPT models (incl. ChatGPT) cannot repeat back certain short strings:
-
Contrasting Bing Search and ChatGPT responses to mistaken math proof request
By
–
An example contrasting Bing Search and ChatGPT responses to a mistaken request for a math proof:
-
Career Advancement Strategies in the LLM Era
By
–
Ready to take your career to the next level in the age of LLMs? Here are some proven tips and strategies to stand out in today's competitive job market! Check out my latest blog to learn more https://
unwindai.substack.com/p/how-to-get-j
ob-ready-in-the-world
…
