If AZ cannot rely on a pre-trained LM to communicate, it is impossible (there is no reason the model learns with self play to communicate with valid sentences). Other issue will of course be the infinite action space (which we also had in theorem proving) but this is manageable
LLMS
-
Large Language Models Could Become Presidents
By
–
If Biden can be president, so can a large language model.
-
LangChain Open Source Demos and Projects Featured on GitHub
By
–
We’re going to assembling a list of open source demos/projects that use LangChain and feature then on our GitHub. Which ones should we include?
-
LangChain Demo Showcasing Prompt Templates LLM Chains
By
–
Very cool demo that is also a great showcase for using LangChain’s prompt templates, LLM wrapper, and chains
-
LLMs as Solutions for Scientific Publication Overload
By
–
“Scientific publication has extended far beyond our ability to keep up" (V Bush, 1945). Could LLM’s help?
https://
arxiv.org/pdf/2211.09085
.pdf
… -
Will Neural Networks Create Novel Symbolic Abstractions?
By
–
When will our large neural network models invent one of these? What will it be?! I dream of this novel knowledge creation, because the generation of symbolic abstractions and descriptions of the world has been great for generalisation beyond the senses.
-
LLM Interactions Beyond Prompts: Integration with Tools and Systems
By
–
Yeah it’s really not limited to prompts talking to eachother, but rather any interactions between an LLM and other things (whether that be another prompt, or a python REPL, or anything)
-
Latest LLM Reads Text on Wall Warning of Impact
By
–
The latest LLM can read writing on a wall. It says "You're about to hit me."
-
LangChain 0.0.21: Memory, Conversational Chain, and Improvements
By
–
LangChain version 0.0.21 Added in Memory abstraction (
@sjwhitmore
) Conversational chain (
@sjwhitmore
) Method for adding examples to ExampleSelector class (
@sjwhitmore
) Bug fix for search class Support for all params in the OpenAI wrapper -
Generative AI framing undersells actual capabilities
By
–
Similar issue with "Generative AI" – so hot right now, but this framing is, believe it or not, *underselling* capabilities we are already seeing, and perhaps causing people to work on the wrong things.