Capabilities and Intelligence Emergence in LLMs
I’ve been processing the new Krakauer, Krakauer & Mitchell paper on emergence in LLMs alongside Michael Levin’s recent work on ingressing mind” and there is much to chew. Both are...
read moreI’ve been processing the new Krakauer, Krakauer & Mitchell paper on emergence in LLMs alongside Michael Levin’s recent work on ingressing mind” and there is much to chew. Both are...
read moreMovement One: The Dance of Understanding I’ve been exploring a fascinating pattern that appears across domains: the principle of “right distance” - not as a fixed measurement, but as a...
read moreWhat does 0-shot CoT Maj@32 means? skip to the end for recap and answer When evaluating the capabilities of LLMs on tasks like question answering, researchers often use benchmarks consisting...
read moreDefinizione di agente Una qualsiasi entità che può percepire il suo ambiente tramite sensori e agire su quell’ambiente tramite attuatori. Nel contesto degli LLM, un agente può essere inteso come...
read morePer migliorare le capacità degli LLM preaddestrati è possibile seguire due strategie: Un approccio basato su Retrieval-Augmented Generation (RAG) Un fine-tuning del modello preaddestrato RAG L’approccio RAG mira a potenziare...
read moreIntroduzione Siamo tutti consapevoli delle potenzialità di ChatGPT. Si tratta probabilmente della killer app della decade (forse del secolo se i progressi continuano allo stesso rate?) e da quando è...
read moreThe new paper on AlphaCodium struck me for its usefulness in crafting good prompts for coding. I built a GPT that tries to emulate the steps. You can find it...
read moreThank you Nikos for the good time in Frankfurt! How numbers are stored in computers the above page also has this must see link Best gitrepo to study and review...
read moreGoogle Dorks from hack5 Lex Fridman conversation with Neri Oxman on silk worms. I did not know that they are not able to fly after metamorphosis, crazy effect of human...
read more