GPT4T retakes the SOTA LLM spot (probably) but also i like this way of announcing things: open source your evals, make it reproduceable and comparable. and **LIGHTWEIGHT**: eleuther’s harness and oai/evals always felt far too heavy of a lift and not flexible. this thing
@swyx
-
InfiniAttention Paper Performance Review for Model Size
By
–
first thing i looked for in the infiniattention paper. theres limits ofc but they do pretty well for the size
-
Google’s Linear Attention Breakthrough: Griffin and Infini-Attention
By
–
with Griffin and Infini-attention, it increasingly feels like Google leapfrogged Together and RWKV in the race for scaling up linear attention, and they shared a watercooler conversation with Anthropic or something papers are trickling out now but it's hard to know what's
-
Emerging Consensus: 8x22B Model is Mixtral Medium
By
–
think emerging consensus is 8x22B -is- mixtral medium
-
Cost Reliability Performance: Flow Engineering and AI Agent Notebooks
By
–
Cost v Reliability v Performance: Choose Three! pod: https://
latent.space/p/elicit
w/ @jungofthewon and @stuhlmueller How do you improve prompt engineering and ai agents? The emerging paradigm is sometimes called "Flow engineering", but @elicitorg has reinvented the Notebook, in -
AI Models: The Risk of Mimicking Human Flaws Too Perfectly
By
–
i feel like there's potential for an Eponymous Law when we make AI model humans but forget that we don't actually want it to be completely human. an RLHF application of the Sour Lesson solutions via @jason_lopatecki and
-
Incredible Prompt Engineering Solution Creating Programming Languages with LLMs
By
–
absolutely incredible solution to the @VictorTaelin challenge (by anime pfp anon ofc) look at this peak prompt engineering llms let you construct entire programming languages from thin air limited only by your imagination
-
LLM as Judge: Are All Results Mediocre 7 out of 10?
By
–
now the bone chilling question for the lazy people who use llm as judge for everything: have you run this analysis on your judging results and is everything ~7 out of 10
-
AI Engineering Manager Enjoying Current Role
By
–
haha yeah. im enjoying being an ai eng manager right now!
-
Inverse Correlation Between Specification Numbers and Model Quality
By
–
inverse correlation between number of numbers on the announcement and the actual quality of model