
Keivan Alizadeh
@KeivanAlizadeh2
Followers
569
Following
218
Media
2
Statuses
33
RT @MFarajtabar: 🧵 1/8 The Illusion of Thinking: Are reasoning models like o1/o3, DeepSeek-R1, and Claude 3.7 Sonnet really "thinking"? 🤔 O….
0
585
0
RT @i_mirzadeh: We have open-sourced GSM-Symbolic templates and generated data! 🎉.- Github: .- Hugging Face: https:….
0
4
0
How we can make the process of RLHF more robust? Using a simple trick: Instead of limiting the KL divergence to a single SFT model we can search around a model soup which resides in a higher reward space. Please check our interns great work!.
1/🔔Excited to share my internship work, SALSA: Soup-based Alignment Learning for Stronger Adaptation, (NeurIPS workshop paper)! 🎉. Proximal Policy Optimization (PPO) often limits exploration by keeping models tethered to a single reference model. SALSA, however, breaks free.
0
0
8
RT @MFarajtabar: 1/ LLM inference is very expensive; and LLMs don't necessarily use their full capacity to respond to a specific prompt. Th….
0
56
0
RT @MFarajtabar: ** Intern position on LLM reasoning **. @mchorton1991, @i_mirzadeh, @KeivanAlizadeh2.and I are co-hosting an intern positi….
0
25
0
RT @andrewglynch: nobody will remember:.- your salary.- how “busy you were”.- how many hours you worked. people will remember:.- nothing. Y….
0
1K
0
RT @MFarajtabar: 1/ Can Large Language Models (LLMs) truly reason? Or are they just sophisticated pattern matchers? In our latest preprint,….
0
1K
0
RT @adityakusupati: Introducting🪆Matryoshka Representations for Adaptive Deployment🪆. TL;DR: up to 14× lower real-world classification & re….
0
99
0
RT @gabriel_ilharco: Instead of a single neural network, why not train lines, curves and simplexes in parameter space?. Fantastic work by @….
0
44
0
RT @gabriel_ilharco: I've been seeing a lot of talk around the recent Vision Transformer (ViT) paper, so I thought I'd highlight some of m….
0
51
0
RT @RAIVNLab: Catch us at this ECCV where @JamesPa91074457 and @sarahmhpratt from our lab present their works as spotlights!!!. VisualCOME….
0
3
0
RT @sacmehtauw: Excited to share our work for diagnosing breast cancer. We extend self-attention mechanism to learn representations on 100s….
0
2
0
Glad to be part of the NED team. A simple framework toward more realistic ML Systems. NED doesn't separate train and test. Just go iN thE wilD, collect data and evaluate. NED extends ML models to ML Systems, which contains both model and training strategy.
Sharing In The Wild: From ML Models to Pragmatic ML systems. In The Wild (NED) is a learning and evaluation framework designed to further progress towards general ML systems capable of excelling in the real world. Paper: Site:
1
3
14
RT @Mitchnw: sharing Supermasks in Superposition (SupSup):. A model that sequentially learns thousands of tasks with negligible forgetting-….
0
56
0
RT @sarahmhpratt: Excited to share Grounded Situation Recognition -- our work (with @yatskar, @LucaWeihs, Ali Farhadi, and @anikembhavi) on….
0
18
0
Check out our work on the Butterfly Transform (BFT), a new building block in convolutional neural networks. BFT fuses information among channels more efficient than standard 1*1 convolutions. PDF: Code: to appear at #CVPR2020
2
29
80
RT @ehsanik: Check out our new work (with Daniel Gordon, Dieter Fox and Ali Farhadi) on unsupervised representation learning from unlabeled….
0
17
0