Ariel Goldstein
@GoldsteinYAriel
Followers
115
Following
88
Media
0
Statuses
48
Joined June 2016
RT @TaubenfeldAmir: New Preprint 🎉. LLM self-assessment unlocks efficient decoding ✅. Our Confidence-Informed Self-Consistency (CISC) metho….
0
20
0
RT @R_Tikochinski: Very excited to share our new paper published in Nature Communications @NatureComms (link below). This work is part of m….
0
3
0
RT @DariaLioub: 🎯 Finally, we leverage these insights to introduce a new token-level early-exit strategy that beats existing methods in bal….
0
1
0
RT @DariaLioub: We propose an underlying task transition mechanism where each task corresponds to determining the k-th ranking token. By un….
0
1
0
RT @DariaLioub: Our findings reveal that this is true across language, vision, and speech models and across architecture variants: decoder….
0
1
0
RT @DariaLioub: 📢Paper release📢.What computation is the Transformer performing in the layers after the top-1 becomes fixed (a so called "sa….
0
14
0
RT @YuvalShalev1: 🧠🤖 How do LLMs think? What kind of thought processes can emerge from artificial intelligence? Our latest paper about mult….
0
8
0
RT @GiliLior: 🧠🤖 Does learning in the brain inherently require plasticity? In our latest paper we question this assumption, by leveraging i….
0
8
0
RT @GabiStanovsky: datasets and tasks, while research in machine consciousness may aim to imbue machines with internal experience or to mea….
0
1
0
RT @GabiStanovsky: supposedly require consciousness, from functional understanding, which measures objective success on given instances. Th….
0
1
0
RT @GabiStanovsky: We show that seminal works in AI would likely consider Z incapable of understanding because, regardless of its external….
0
1
0
RT @GabiStanovsky: To demonstrate this, imagine a “zombie” chatbot Z. It is completely open-source (à la OLMo), and it excels in all curren….
0
1
0
RT @GabiStanovsky: 🧠 Can LLMs truly understand text, or are they mere 🦜? We argue that this debate implicitly revolves around the function….
0
9
0
RT @TaubenfeldAmir: How do the inherent biases in Large Language Models impact their capacity for accurately simulating human behavior?. We….
arxiv.org
The emergence of Large Language Models (LLMs), has opened exciting possibilities for constructing computational simulations designed to replicate human behavior accurately. Current research...
0
4
0
RT @GoldsteinAlon_: 1/9: Our recent study, Decoding Stumpers: Large Language Models vs. Human Problem-Solvers, was published in #EMNLP2023.….
0
5
0
RT @EricHam4: In a new version of our (@marianoschain, @ArielYGoldstein, @HassonLab) paper ( we show evidence that….
arxiv.org
Deep Language Models (DLMs) provide a novel computational paradigm for understanding the mechanisms of natural language processing in the human brain. Unlike traditional psycholinguistic models,...
0
8
0
RT @NeumanRoee: זה הסלפי של אחרי העברת חוק הסבירות. באותה שעה בכנסת עמדו אלופים ביחד עם הרמטכ״ל בכנסת לאחר שהתחננו לדבר איתם, להסביר להם שמ….
0
1K
0
RT @HassonLab: We're posting this preprint alongside another @HassonLab manuscript led by @zaidzada_ and @samnastase where we use large lan….
0
2
0