Tired of using an "unofficial" ChatGPT api? Just use this instead! More seriously: applications like ChatGPT are very possible for anyone to develop NOW. What will you build?
LLMS
-
LangChain Already Supports Conversation Summary Memory
By
–
we actually support that already! https://
langchain.readthedocs.io/en/latest/gett
ing_started/memory.html#ConversationChain-with-ConversationSummaryMemory
… -
Dynamic conversation context length determination in LangChain
By
–
excited to hear your thoughts/ideas! right now its pretty simple: just take the last N lines of conversation. but a very natural extension to determine N dynamically based on their length
-
Guidance-Annealing Approach Combined with Normalization Techniques
By
–
Have you tried the guidance-annealing approach too? It works well combined with this normalization.
-
Discovering Latent Knowledge in Language Models Without Supervision
By
–
Discovering Latent Knowledge in Language Models Without Supervision Burns et al.: https://
arxiv.org/abs/2212.03827 #ArtificialIntelligence #DeepLearning #MachineLearning -
Elixir Now Supports AI Models from GPT2 to Stable Diffusion
By
–
welcomes @elixirlang ! Elixir now supports models on the Hub, from GPT2 to Stable Diffusion We're excited to see what the community builds with this! @josevalim
, the creator of Elixir, wrote this nice blog post -
Deep Learning Breakthroughs Make Research Agendas Obsolete Quickly
By
–
Maybe one of the most bitter lesson for researchers in DL/ML: an entire research agenda can be rendered moot in a matter of weeks
-
Text-davinci-003 results now available in paper appendix
By
–
https://
arxiv.org/abs/2210.11416 Appendix has text-davinci-003 results now -
LLM Text Summarization More Valuable Than Expansion
By
–
people seem very excited that chatgpt can expand a few bullet points into a lot of well-written text. it can also collapse a lot of well-written text into a a few bullet points. the latter is much more valuable, right?
-
Recent papers on LLMs for formal mathematics and theorem proving
By
–
We had two papers on this topic recently, one using a MCTS approach on formal systems (such as Lean or Metamath): https://
arxiv.org/abs/2205.11491 and one where we combine LLMs with formal provers to verify the generations: https://
arxiv.org/abs/2210.12283