ChatGPT is incredibly limited, but good enough at some things to create a misleading impression of greatness. it's a mistake to be relying on it for anything important right now. it’s a preview of progress; we have lots of work to do on robustness and truthfulness.
LLMS
-
The Evolution of Reality: From Physical to Digital Definitions
By
–
The evolving definition of reality:
1980s: Reality.
2000s: Search engine results.
2020s: Chatbot output. -
Chain-of-thought Jar Jar Binks posts linked by Twitter algorithm
By
–
I've never seen Twitter's timeline algorithm more excited than when, immediately after I posted about chain-of-thought prompting GPT‑3 in the style of Jar Jar Binks, it showed me @paulg
's tweet about the chain of thought that prompted Jar Jar Binks. -
ChatGPT smarter than Davinci 002; Davinci 003 smarter than both
By
–
I think ChatGPT is smarter than Davinci 002, though less compliant. Davinci 003 is smarter than both, at least for the important use cases:
-
LLMs Fact-Checking Themselves Through Prompt Chaining
By
–
@0xjasper came up with a very cool example of LLMs fact checking themselves with prompt chaining. See the original tweet thread below
-
LangChain version 0.0.34: OpenAI moderation and fact checking
By
–
LangChain version 0.0.34 Doc cleanup from `moozilla` (if on twitter lmk!) Support for @OpenAI's moderation endpoint LLMFactChecker chain from @thepromptking, original idea by @0xjasper Some more info on the LLM fact checker chain
-
ChatGPT generates plausible but incorrect answers without online feedback
By
–
I spoke to @latimes on #ChatGPT on how it can confidently give wrong but plausible answers to puzzles, and it is not getting online feedback on if the text it is generating is true.
-
Simple conversation buffer memory with dynamic N adjustment
By
–
it is pretty simple at the moment – just keeps a buffer of the past N conversations in memory. a pretty simple improvement would be dynamically adjusting N based on the length of the conversations
-
Improving ChatGPT’s Math Abilities with LangChain
By
–
An awesome example improving ChatGPT's much maligned math abilities!