New paper — In ‘Unnatural Instructions’ we propose a new method to automatically generate natural language instructions, allowing us to scale up to 240K diverse instructions & train models that rival the performance of contemporary instruction-tuned models. Read on ArXiv
RESEARCH
-
Groq First Vendor with Published Audited AI Benchmarking Results
By
–
Peter Nabicht, President of STAC – Securities Technology Analysis Center, announced that @GroqInc is the first vendor to provide the financial services industry with published audited benchmarking results, making real-time #AI more accessible. Read more at http://
groq.link/stacpr -
Award-Winning Research on Language Grounding in Robotic Affordances
By
–
Congratulations to the authors of “Do As I Can, Not As I Say: Grounding Language in Robotic Affordances” (see the blog at https://
goo.gle/3QRJhgl) for winning the Special Innovation Award at @corl_conf
! #CoRL2022 -
LLMs Reliability Limitations and Twitter Indexing Accuracy
By
–
LLMs aren't yet 100% reliable, and our Twitter indexing isn't exhaustive. So Bird SQL may not be accurate on all queries.
— Perplexity (@perplexity_ai) 20 décembre 2022
Help us improve by sharing suggestions and feedback via Twitter, Discord (https://t.co/R4G21AmwQ7), support@perplexity.ai
Also seehttps://t.co/wgT9djadYnLLMs aren't yet 100% reliable, and our Twitter indexing isn't exhaustive. So Bird SQL may not be accurate on all queries. Help us improve by sharing suggestions and feedback via Twitter, Discord (
https://
discord.com/invite/kWJZsxP
DuX
…), support@perplexity.ai Also see -
Quantum advantage breakthrough achievement announced
By
–
Sounds like a super cool quantum advantage result!
-
AGI Debate: World’s Distinguished Researchers Panel Discussion
By
–
AGI DEBATE "The Debate The World Is Waiting For" Fri, Dec 23, 2022 Website: https://
agidebate.com It takes a village . . . Four panels of the world's most distinguished researchers and experts on AGI. #AGI #AIDebate #AGIDebate -
MMR Example Selector for Optimal Contextual Diversity
By
–
Max Marginal Relevance (MMR) Example Selector Based on this preprint by @xiye_nlp et al, the MMR example selector chooses diverse, in-context examples H/t @johnjnay for surfacing this paper Docs: https://langchain.readthedocs.io/en/latest/examples/prompts/prompt_management.html Original Paper: https://x.com/xiye_nlp/status/1603821850592628738?s=20&t=wQPcRd8IJ8URbmytuYP7QA
-
Gradient Backpropagation and Frequent Word Prioritization in Training
By
–
Because we sum, the gradients for each token get backpropagated to all the rows that were used for it. So with enough training, the model ends up at a good compromise. Frequent words are naturally prioritised in this, because they'll simply have more gradients.
-
Hash Embedding Performance on Small Datasets Analysis
By
–
However, the hash embedding doesn't do so well if the datasets are too small. That's one of the results that I found really useful in this tech report. It's great to have this all explored more thoroughly, and we look forward to publishing more experiments like this in future.
-
spaCy’s Hashing Approach for Word Representations Learning
By
–
This is almost what we do in spaCy — it's just that, instead of reserving the first 9900 rows for common words, we just use the hashing. Given that the assignment of words to shared rows is arbitrary, how can the model learn good representations?