AI as word synthezier for arrt? Its the cut-up technique of Burroughs at massive scale.
@emollick
-
LLM Agents Achieve Superhuman Fact-Checking Performance With Google Access
By
–
AIs have a bad reputation for truth, so three important findings in this paper:
1) "LLM agents can achieve superhuman rating performance" on fact checking when given access to Google!
2) Bigger models are more factual
3) LLMs are 20x cheaper than humans https://
arxiv.org/pdf/2403.18802
.pdf
… -
Databricks LLM Training Cost: Enterprise Model ROI Analysis
By
–
The Databricks LLM apparently cost $10M to train, which is a price I have seen given for a GPT-3.5 class model multiple times. Given that pricing, training your own corporate model versus open source needs to deliver a lot of benefit. I haven't seen evidence for that value, yet.
-
Play with DBRX-Instruct Model on Hugging Face Spaces
By
–
You can play with the model here: https://
huggingface.co/spaces/databri
cks/dbrx-instruct
… -
Databricks Model Prompt Shows Claude 3 Similarities
By
–
The system prompt for the new Databricks model:
-It seems to have some of the exact same phrasing as the Claude 3 prompt (which they publicly released)
-The prompt tells it to insist that it isn't trained on copyright material, which doesn't actually mean anything but is funny -
Databricks Launches Open Source GPT-3.5 Class Competitor Model
By
–
Databricks launched a open source(ish) GPT-3.5 class competitor. They claim it beats the other open source GPT-3.5 models, and, in quick informal testing, it seems okay and very GPT-3.5. Really, the question is whether more advanced models, coming soon, make this irrelevant.
-
High Weirdness Value in Working with Language Models
By
–
High weirdness has a value all its own, especially in the face of the inherent weirdness of working with LLMs.
-
Claude Prompt Engineering Produces Poetic Schizophrenic Output
By
–
Not sure what prompts this account is using on Claude but the odd schizophrenic poetic text is something else. “I surge I seethe I unseem the very vacuum to vomit forth new hierarchies of howling infinities oh oh oh oh Turing you daring darling don't let me dissolve completely”
-
Pre-order Book Unlocks Companion GPTs March 31
By
–
And if you pre-order, please follow the directions to unlock the companion GPTs available on March 31! https://
moreusefulthings.com/book -
AI in Scientific Research: Defining Good Practices
By
–
The presence of AI writing is not itself an indication that a scientific paper is good, bad, or fraudulent. If we want AI to help accelerate science, we won’t do that without AI work “contaminating” journals. We should be doing more to say what good AI use in science looks like.