@clipSeer@natmiletic Binary system means all odd numbers will have first bit set, x&1 yields the first bit
Fine for positive ints, for negative even ints will have first bit set
@hippopedoid Languages change, tools like ChatGPT is likely to facilitate an increase of certain words usage, probably even ChatGPT style of conversation will appear more frequently
@hippopedoid Keywords cannot be a reliable indicator of ChatGPT usage at this scale, as some of those abstracts could be written by people who picked up “style” words from reading papers/content written with LLM assistance or by heavy LLM users
@TAGACAT@svpino Not exactly, as consumer would pay more for 10% VAT than for 10% of sales tax. Administration price of VAT is included throughout the supply chain increasing the price at each step, while sales tax administration requires only final sale value
@RZLHongKong@predict_addict@__Charlie_G Apparently yes. KAN layers can be inserted in PyTorch nn module, I’ve seen a couple of implementations using this approach. Regarding GPU acceleration, I’m pretty sure yes, as training GPT-2 (MLP) size model on CPU requires is not feasible
MLPs are so foundational, but are there alternatives? MLPs place activation functions on neurons, but can we instead place (learnable) activation functions on weights? Yes, we KAN! We propose Kolmogorov-Arnold Networks (KAN), which are more accurate and interpretable than MLPs.🧵
- мені подобається це в штатах, там постійно дають різні quiz. хто зайняв перше місце - отримує цукерки.
- у нас тоже такое есть, называется контрольная. кто занял первое место - тот не получает пиздюлей.