Yes it'd be interesting to see the 8B model in 8-bit precision vs. the 4B model in BF16.
RESEARCH
-
Phi-2 vs Phi-3: Model Comparison and Benchmark Relevance
By
–
I think they just want to compare it to the 4B model. But I agree, phi-2 isn't the best comparison point. It's funny it's still regularly used in benchmarks despite the release of phi-3.
-
Teaching AI: Pedagogy, Materials, and Neutrality Questions
By
–
But we're talking about the *teaching* of it. What do you teach? How? To whom? With what materials? These are obviously not inherently neutral.
-
Content Curation and Bias in AI Material Presentation
By
–
And of course what is and isn't in that material, and how its presented, isn't something that just happened by chance.
-
Attention Mechanism Importance in AI Development
By
–
Attention is all you need (to the y axis in this case)
-
AI Revolution: Hype or Genuine Industrial Transformation Impact?
By
–
AI: Overhyped Fantasy Or Truly The Next Industrial Revolution? Delve into the debates around #AI's #revolutionary impact on #society, work, and #ethical boundaries, and contemplate its role in shaping our #future.
-
ContinualAI Unconference Proceedings Volume 249 Now Available PMLR
By
–
Volume 249 Proceedings of the 1st ContinualAI Unconference https://
proceedings.mlr.press/v249 Is now available on PMLR -
Teaching AI knowledge requires engaging with its political context
By
–
Yes, but that doesn't mean "Teaching this knowledge requires you to engage with its politics" is incorrect. If you choose to ignore all the context and history behind an idea when teaching it, you *are* engaging in politics in that choice.
-
Generative Photomontage: Controlling Text-to-Image Generation
By
–
Generative Photomontage discussion: https://
huggingface.co/papers/2408.07
116
… Text-to-image models are powerful tools for image creation. However, the generation process is akin to a dice roll and makes it difficult to achieve a single image that captures everything a user wants. In this paper, -
InfinityMATH: Scalable Dataset for Mathematical Reasoning
By
–
InfinityMATH A Scalable Instruction Tuning Dataset in Programmatic Mathematical Reasoning discussion: https://
huggingface.co/papers/2408.07
089
… Recent advancements in Chain-of-Thoughts (CoT) and Program-of-Thoughts (PoT) methods have greatly enhanced language models' mathematical reasoning