Split (1)

train · 587 rows

filename stringlengths 9 127	text stringlengths 133 11k
2310.15154.pdf	Pre-publication draft LINEAR REPRESENTATIONS OF SENTIMENT INLARGE LANGUAGE MODELS Curt Tigges♣, Oskar John Hollinsworth♡, Atticus Geiger♠⋆, Neel Nanda♢ ♣EleutherAI Institute,♡SERI MATS,♠Stanford University,⋆Pr(Ai)2R Group,♢Independent *Equal primary authors (order random) ABSTRACT Sentiment is a pervasive feature in ...
2212.10559.pdf	Why Can GPT Learn In-Context? Language Models Secretly Perform Gradient Descent as Meta-Optimizers Damai Dai†∗, Yutao Sun∥∗, Li Dong‡, Yaru Hao‡, Zhifang Sui†, Furu Wei‡ †Peking University∥Tsinghua University ‡Microsoft Research https://github.com/microsoft/LMOps Abstract Large pretrained language models have shown sur...
2306.00297.pdf	Transformers learn to implement preconditioned gradient descent for in-context learning Kwangjun Ahn1,3,, Xiang Cheng1,3,, Hadi Daneshmand2,3,*, and Suvrit Sra1,3 1Department of Electrical Engineering and Computer Science, MIT 2Foundations of Data Science Institute (FODSI) 3Laboratory for Information and Decision Sys...
2105.14368.pdf	Fit without fear: remarkable mathematical phenomena of deep learning through the prism of interpolation Mikhail Belkin Halicio˘ glu Data Science Institute, University of California San Diego La Jolla, USA In memory of Partha Niyogi, a thinker, a teacher, and a dear friend. Abstract In the past decade the mathematical t...
2306.09927.pdf	arXiv:2306.09927v1 [stat.ML] 16 Jun 2023Trained Transformers Learn Linear Models In-Context Ruiqi Zhang UC Berkeley rqzhang@berkeley.eduSpencer Frei UC Berkeley frei@berkeley.edu Peter L. Bartlett UC Berkeley and Google DeepMind peter@berkeley.edu June 19, 2023 Abstract Attention-based neural networks such as transfo...
2310.15418.pdf	Fractal Landscapes in Policy Optimization Tao Wang UC San Diego taw003@ucsd.eduSylvia Herbert UC San Diego sherbert@ucsd.eduSicun Gao UC San Diego sicung@ucsd.edu Abstract Policy gradient lies at the core of deep reinforcement learning (RL) in continuous domains. Despite much success, it is often observed in practice t...
2205.14135.pdf	FlashAttention : Fast and Memory-Eﬃcient Exact Attention with IO-Awareness Tri Daoy, Daniel Y. Fuy, Stefano Ermony, Atri Rudraz, and Christopher Réy yDepartment of Computer Science, Stanford University zDepartment of Computer Science and Engineering, University at Buﬀalo, SUNY {trid,danfu}@cs.stanford.edu ,ermon@stanfo...
bayesian-interactive-optimization.pdf	Eurographics/ ACM SIGGRAPH Symposium on Computer Animation (2010) M. Otaduy and Z. Popovic (Editors) A Bayesian Interactive Optimization Approach to Procedural Animation Design Eric Brochu Tyson Brochu Nando de Freitas University of British Columbia Abstract The computer graphics and animation ﬁelds are ﬁlled with appl...
Introduction to Probabilistic Topic Models.pdf	Introduction to Probabilistic Topic Models David M. Blei Princeton University Abstract Probabilistic topic models are a suite of algorithms whose aim is to discover the hidden thematic structure in large archives of documents. In this article, we review the main ideas of this ﬁeld, survey the current state-of-the-art, ...
GPT-2.pdf	Language Models are Unsupervised Multitask Learners Alec Radford1Jeffrey Wu1Rewon Child1David Luan1Dario Amodei1Ilya Sutskever1 Abstract Natural language processing tasks, such as ques- tion answering, machine translation, reading com- prehension, and summarization, are typically approached with supervised learni...
2647-elbo-ing-stein-mixtures.pdf	Under review as a conference paper at ICLR 2023 ELBO- INGSTEIN MIXTURES Anonymous authors Paper under double-blind review ABSTRACT Stein variational gradient descent (SVGD) (Liu & Wang, 2016) is a particle-based technique for Bayesian inference. SVGD has recently gained popularity because it combines the ability of var...
1711.00165.pdf	Published as a conference paper at ICLR 2018 DEEPNEURAL NETWORKS AS GAUSSIAN PROCESSES Jaehoon Lee∗†, Yasaman Bahri∗†, Roman Novak , Samuel S. Schoenholz, Jeffrey Pennington, Jascha Sohl-Dickstein Google Brain {jaehlee, yasamanb, romann, schsam, jpennin, jaschasd }@google.com ABSTRACT It has long been known that a sing...
2210.03370.pdf	GNM: A General Navigation Model to Drive Any Robot Dhruv Shah†β, Ajay Sridhar†β, Arjun Bhorkarβ, Noriaki Hiroseβτ, Sergey Levineβ 𝜏 ot og GNM Training Large Heterogeneous Datasets Fig. 1: A general navigation model to drive any robot. By training on diverse, heterogeneous datasets, a single “omnipolicy” can contro...
2203.03466.pdf	Tensor Programs V: Tuning Large Neural Networks via Zero-Shot Hyperparameter Transfer Greg Yang∗×Edward J. Hu∗×†Igor Babuschkin◦Szymon Sidor◦Xiaodong Liu× David Farhi◦Nick Ryder◦Jakub Pachocki◦Weizhu Chen×Jianfeng Gao× ×Microsoft Corporation◦OpenAI Abstract Hyperparameter (HP) tuning in deep learning is an expensive pr...
1705.01509.pdf	Neural Models for Information Retrieval Bhaskar Mitra Microsoft, UCL∗ Cambridge, UK bmitra@microsoft.comNick Craswell Microsoft Bellevue, USA nickcr@microsoft.com Abstract Neural ranking models for information retrieval (IR) use shallow or deep neural networks to rank search results in response to a query. Traditional ...
2312.12456.pdf	arXiv:2312.12456v1 [cs.LG] 16 Dec 2023PowerInfer: Fast Large Language Model Serving with a Consum er-grade GPU Yixin Song, Zeyu Mi∗, Haotong Xie and Haibo Chen Institute of Parallel and Distributed Systems (IPADS), Sha nghai Jiao Tong University Abstract This paper introduces PowerInfer, a high-speed Large Lan- guage...
1802.09568.pdf	arXiv:1802.09568v2 [cs.LG] 2 Mar 2018Shampoo: Preconditioned Stochastic Tensor Optimization Vineet Gupta6Tomer Koren6Yoram Singer‹ March 5, 2018 Abstract Preconditioned gradient methods are among the most general and powerful tools in optimization. However, preconditioning requires storing and manipulating prohibitiv...
Variational Inference.pdf	Variational Inference David M. Blei 1 Set up •As usual, we will assume that x=x1:nare observations and z=z1:mare hidden variables. We assume additional parameters αthat are ﬁxed. •Note we are general—the hidden variables might include the “parameters,” e.g., in a traditional inference setting. (In that case, αare the h...
2208.02813.pdf	Towards Understanding Mixture of Experts in Deep Learning Zixiang Chen∗and Yihe Deng†and Yue Wu‡and Quanquan Gu§and Yuanzhi Li¶ Abstract The Mixture-of-Experts (MoE) layer, a sparsely-activated model controlled by a router, has achieved great success in deep learning. However, the understanding of such architecture rem...
wenzel20a.pdf	How Good is the Bayes Posterior in Deep Neural Networks Really? Florian Wenzel* 1Kevin Roth* + 2Bastiaan S. Veeling* + 3 1Jakub ´Swi ˛ atkowski4 +Linh Tran5 + Stephan Mandt6 +Jasper Snoek1Tim Salimans1Rodolphe Jenatton1Sebastian Nowozin7 + Abstract During the past ﬁve years the Bayesian deep learn- ing community has de...
2301.13856.pdf	Simplex Random Features Isaac Reid1Krzysztof Choromanski* 2 3Valerii Likhosherstov1Adrian Weller* 1 4 Abstract We present Simplex Random Features (SimRFs), a new random feature (RF) mechanism for unbi- ased approximation of the softmax and Gaussian kernels by geometrical correlation of random pro- jection vectors. We p...
2024.02.06.579080v1.full.pdf	Direct Coupling Analysis and the Attention Mechanism 1 Francesco Caredda1†and Andrea Pagnani1,2,3† 2 1DISAT, Politecnico di Torino, Corso Duca degli Abruzzi, 24, I-10129, Torino, Italy 3 2Italian Institute for Genomic Medicine, IRCCS Candiolo, SP-142, I-10060, 4 Candiolo, Italy 5 3INFN, Sezione di Torino, Torino, Via P...
2307.08691.pdf	FlashAttention-2 : Faster Attention with Better Parallelism and Work Partitioning Tri Dao1,2 1Department of Computer Science, Princeton University 2Department of Computer Science, Stanford University trid@cs.stanford.edu July 18, 2023 Abstract Scaling Transformers to longer sequence lengths has been a major problem in ...
supplementary-gpsa.pdf	Supplementary Information for: Generative Capacity of Probabilistic Protein Sequence Models Francisco McGee Sandro Hauri Quentin Novinger Slobodan Vucetic Ronald M. Levy Vincenzo Carnevale Allan Haldane Supplementary Note 1 - sVAE implementation The standard variational autoencoder (sVAE) is a deep, symmetrical, and un...
2010.02502.pdf	Published as a conference paper at ICLR 2021 DENOISING DIFFUSION IMPLICIT MODELS Jiaming Song, Chenlin Meng & Stefano Ermon Stanford University {tsong,chenlin,ermon }@cs.stanford.edu ABSTRACT Denoising diffusion probabilistic models (DDPMs) have achieved high qual- ity image generation without adversarial training, yet...
1901.09321.pdf	Published as a conference paper at ICLR 2019 FIXUP INITIALIZATION : RESIDUAL LEARNING WITHOUT NORMALIZATION Hongyi Zhang∗ MIT hongyiz@mit.eduYann N. Dauphin† Google Brain yann@dauphin.ioTengyu Ma‡ Stanford University tengyuma@stanford.edu ABSTRACT Normalization layers are a staple in state-of-the-art deep neural networ...
2402.03300.pdf	DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models Zhihong Shao1,2∗†, Peiyi Wang1,3∗†, Qihao Zhu1,3∗†, Runxin Xu1, Junxiao Song1 Mingchuan Zhang1, Y.K. Li1, Y. Wu1, Daya Guo1∗ 1DeepSeek-AI,2Tsinghua University,3Peking University {zhihongshao,wangpeiyi,zhuqh,guoday}@deepseek.com https://g...
2111.02080.pdf	An Explanation of In-context Learning as Implicit Bayesian Inference Sang Michael Xie Stanford University xie@cs.stanford.eduAditi Raghunathan Stanford University aditir@stanford.edu Percy Liang Stanford University pliang@cs.stanford.eduTengyu Ma Stanford University tengyuma@cs.stanford.edu Abstract Large language mode...
2404.16014v1.pdf	2024-4-25 Improving Dictionary Learning with Gated Sparse Autoencoders Senthooran Rajamanoharan, Arthur Conmy, Lewis Smith, Tom Lieberum†, Vikrant Varma†, János Kramár, Rohin Shah and Neel Nanda *: Joint contribution.†: Core infrastructure contributor. Recent work has found that sparse autoencoders (SAEs) are an effe...
1811.07871.pdf	Scalable agent alignment via reward modeling: a research direction Jan Leike DeepMindDavid Krueger∗ DeepMind MilaTom Everitt DeepMindMiljan Martic DeepMindVishal Maini DeepMindShane Legg DeepMind Abstract One obstacle to applying reinforcement learning algorithms to real-world problems is the lack of suitable reward fu...
1803.03635.pdf	Published as a conference paper at ICLR 2019 THELOTTERY TICKET HYPOTHESIS : FINDING SPARSE , TRAINABLE NEURAL NETWORKS Jonathan Frankle MIT CSAIL jfrankle@csail.mit.eduMichael Carbin MIT CSAIL mcarbin@csail.mit.edu ABSTRACT Neural network pruning techniques can reduce the parameter counts of trained net- works by over ...
2002.10957v2.pdf	MINILM: Deep Self-Attention Distillation for Task-Agnostic Compression of Pre-Trained Transformers Wenhui Wang Furu Wei Li Dong Hangbo Bao Nan Yang Ming Zhou Microsoft Research {wenwan,fuwei,lidong1,t-habao,nanya,mingzhou}@microsoft.com Abstract Pre-trained language models (e.g., BERT (Devlin et al., 2018) and its vari...
reka-vibe-eval.pdf	Vibe-Eval: A hard evaluation suite for measuring progress of multimodal language models Piotr Padlewski∗Max Bain∗Matthew Henderson Zhongkai Zhu Nishant Relan Hai Pham Donovan Ong Kaloyan Aleksiev Aitor Ormazabal Samuel Phua Ethan Yeo Eugenie Lamprecht Qi Liu Yuqi Wang Eric Chen Deyu Fu Lei Li Che Zheng Cyprien de Masso...
10.1016.j.cell.2023.12.012.pdf	Article Human fetal brain self-organizes into long-term expanding organoids Graphical abstract Highlights dHuman fetal brain organoids (FeBOs) display cellular heterogeneity and can be expanded dFeBOs produce a tissue-like ECM niche and enable ECMperturbation studies dDerivation of regional FeBOs allows the study of re...
2009.01325v3.pdf	Learning to summarize from human feedback Nisan Stiennon∗Long Ouyang∗Jeff Wu∗Daniel M. Ziegler∗Ryan Lowe∗ Chelsea Voss∗Alec Radford Dario Amodei Paul Christiano∗ OpenAI Abstract As language models become more powerful, training and evaluation are increas- ingly bottlenecked by the data and metrics used for a particular...
2306.04751.pdf	How Far Can Camels Go? Exploring the State of Instruction Tuning on Open Resources Yizhong Wang∗♣♠Hamish Ivison∗♣Pradeep Dasigi♣Jack Hessel♣ Tushar Khot♣Khyathi Raghavi Chandu♣David Wadden♣Kelsey MacMillan♣ Noah A. Smith♣♠Iz Beltagy♣Hannaneh Hajishirzi♣♠ ♣Allen Institute for AI♠University of Washington {yizhongw,hamish...
2403.19887.pdf	Jamba: A Hybrid Transformer-Mamba Language Model Opher Lieber∗Barak Lenz∗Hofit Bata Gal Cohen Jhonathan Osin Itay Dalmedigos Erez Safahi Shaked Meirom Yonatan Belinkov Shai Shalev-Shwartz Omri Abend Raz Alon Tomer Asida Amir Bergman Roman Glozman Michael Gokhman Avashalom Manevich Nir Ratner Noam Rozen Erez Shwartz Mor...
20-302.pdf	Journal of Machine Learning Research 22 (2021) 1-35 Submitted 3/20; Revised 10/20; Published 3/21 Attention is Turing Complete Jorge P´ erez jperez@dcc.uchile.cl Department of Computer Science Universidad de Chile IMFD Chile Pablo Barcel´ o pbarcelo@uc.cl Institute for Mathematical and Computational Engineering School ...
Pretrained Transformers for Text Ranking: BERT and Beyond.pdf	Pretrained Transformers for Text Ranking: BERT and Beyond Jimmy Lin,1Rodrigo Nogueira,1and Andrew Yates2,3 1David R. Cheriton School of Computer Science, University of Waterloo 2University of Amsterdam 3Max Planck Institute for Informatics Version 0.99 — August 20, 2021 Abstract The goal of text ranking is to generate ...
2212.10560.pdf	SELF-INSTRUCT : Aligning Language Model with Self Generated Instructions Yizhong Wang♣Yeganeh Kordi♢Swaroop Mishra♡Alisa Liu♣ Noah A. Smith♣+Daniel Khashabi♠Hannaneh Hajishirzi♣+ ♣University of Washington♢Tehran Polytechnic♡Arizona State University ♠Johns Hopkins University+Allen Institute for AI yizhongw@cs.washington...
2310.18313.pdf	FP8-LM: Training FP8 Large Language Models Houwen Peng∗Kan Wu∗Yixuan Wei∗ Guoshuai Zhao Yuxiang Yang Ze Liu Yifan Xiong Ziyue Yang Bolin Ni Jingcheng Hu Ruihang Li Miaosen Zhang Chen Li Jia Ning Ruizhe Wang Zheng Zhang Shuguang Liu Joe Chau Han Hu†Peng Cheng† Microsoft Azure and Microsoft Research Abstract In this pape...
2307.10169.pdf	Challenges and Applications of Large Language Models Jean Kaddourα,†,∗, Joshua Harrisβ,∗, Maximilian Mozesα, Herbie Bradleyγ,δ,ϵ, Roberta Raileanuζ, and Robert McHardyη,∗ αUniversity College LondonβUK Health Security AgencyγEleutherAI δUniversity of CambridgeϵStability AIζMeta AI ResearchηInstaDeep Abstract Large Langu...
2401.01325.pdf	LLM Maybe LongLM: Self-Extend LLM Context Window Without Tuning Hongye Jin1 Xiaotian Han1 Jingfeng Yang2Zhimeng Jiang1Zirui Liu3Chia-Yuan Chang1 Huiyuan Chen4Xia Hu3 Abstract This work elicits LLMs’ inherent ability to handle long contexts without fine-tuning. The limited length of the training sequence during traini...
2304.11082.pdf	Preprint. Under review. FUNDAMENTAL LIMITATIONS OF ALIGNMENT INLARGE LANGUAGE MODELS Yotam Wolf∗ The Hebrew University yotam.wolf@cs.huji.ac.ilNoam Wies∗ The Hebrew University noam.wies@cs.huji.ac.il Yoav Levine AI21 Labs yoavl@ai21.comAmnon Shashua The Hebrew University shashua@cs.huji.ac.il ABSTRACT An important aspe...
2302.04065.pdf	Monge, Bregman and Occam: Interpretable Optimal Transport in High-Dimensions with Feature-Sparse Maps Marco Cuturi1Michal Klein1Pierre Ablin1 Abstract Optimal transport (OT) theory focuses, among all mapsT:Rd→Rdthat can morph a prob- ability measure onto another, on those that are the “thriftiest”, i.e. such that the a...
2106.09685.pdf	LORA: L OW-RANK ADAPTATION OF LARGE LAN- GUAGE MODELS Edward Hu∗Yelong Shen∗Phillip Wallis Zeyuan Allen-Zhu Yuanzhi Li Shean Wang Lu Wang Weizhu Chen Microsoft Corporation {edwardhu, yeshe, phwallis, zeyuana, yuanzhil, swang, luw, wzchen }@microsoft.com yuanzhil@andrew.cmu.edu (Version 2) ABSTRACT An important paradigm...
10.2307.2334029.pdf	A note on DPO with noisy preferences & relationship to IPO Eric Mitchell November 25, 2023 (v1.1) ‘OG’ RLHF aims for reward maximization with a KL constraint to reference model πref(inputs xomitted): π∗= argmax πEy∼π r(y)−βlogπ(y) πref(y) (1) DPO [3] derives a loss on the current policy πθ(where our dataset says ywis...
10.1016.j.cell.2024.01.003.pdf	Leading Edge Commentary Structure is beauty, but not always truth James S. Fraser1,and Mark A. Murcko2, 1Department of Bioengineering and Therapeutic Sciences, University of California San Francisco, San Francisco, CA, USA 2Disruptive Biomedical LLC, Holliston, MA, USA *Correspondence: jfraser@fraserlab.com (J.S.F.),...
2307.13304.pdf	QuIP: 2-Bit Quantization of Large Language Models With Guarantees Jerry Chee Department of Computer Science Cornell University jerrychee@cs.cornell.eduYaohui Cai Department of Electrical and Computer Engineering Cornell University yc2632@cornell.edu Volodymyr Kuleshov Department of Computer Science Cornell University k...
2203.02155.pdf	Training language models to follow instructions with human feedback Long Ouyang∗Jeff Wu∗Xu Jiang∗Diogo Almeida∗Carroll L. Wainwright∗ Pamela Mishkin∗Chong Zhang Sandhini Agarwal Katarina Slama Alex Ray John Schulman Jacob Hilton Fraser Kelton Luke Miller Maddie Simens Amanda Askell†Peter Welinder Paul Christiano∗† Jan ...
2305.14992.pdf	Reasoning with Language Model is Planning with World Model Shibo Hao∗♣Yi Gu∗ ∗♣Haodi Ma♢Joshua Jiahua Hong♣Zhen Wang♣ ♠ Daisy Zhe Wang♢Zhiting Hu♣ ♣UC San Diego,♢University of Florida ♠Mohamed bin Zayed University of Artificial Intelligence {s5hao, yig025, jjhong, zhw085, zhh019}@ucsd.edu {ma.haodi, daisyw}@ufl.edu Abs...
1606.06565.pdf	Concrete Problems in AI Safety Dario Amodei∗ Google BrainChris Olah∗ Google BrainJacob Steinhardt Stanford UniversityPaul Christiano UC Berkeley John Schulman OpenAIDan Man´ e Google Brain Abstract Rapid progress in machine learning and artiﬁcial intelligence (AI) has brought increasing atten- tion to the potential imp...
10.1016.j.cell.2023.12.010.pdf	Article Hypoxia and intra-complex genetic suppressors rescue complex I mutants by a shared mechanism Graphical abstract Highlights dHypoxia rescue and hyperoxia sensitivity of complex I mutants are conserved in C. elegans dHypoxia rescue is independent of HIF activation or attenuation of ROS toxicity dNDUFA6/nuo-3(G60D...
2402.04362v2.pdf	Neural Networks Learn Statistics of Increasing Complexity Nora Belrose1Quintin Pope2Lucia Quirke1Alex Mallen1Xiaoli Fern2 Abstract The distributional simplicity bias (DSB) posits that neural networks learn low-order moments of the data distribution first, before moving on to higher-order correlations. In this work, we ...
2210.05845.pdf	Contrastive Retrospection: honing in on critical steps for rapid learning and generalization in RL Chen Sun∗ Mila, Université de Montréal sunchipsster@gmail.comWannan Yang New York University winnieyangwn96@gmail.comThomas Jiralerspong Mila, Université de Montréal thomas.jiralerspong @mila.quebec Dane Malenfant McGill ...
2311.11829.pdf	System 2 Attention (is something you might need too) Jason Weston MetaSainbayar Sukhbaatar Meta Abstract Soft attention in Transformer-based Large Language Models (LLMs) is sus- ceptible to incorporating irrelevant information from the context into its latent representations, which adversely affects next token generati...
2404.10642v1.pdf	Self-playing Adversarial Language Game Enhances LLM Reasoning Pengyu Cheng, Tianhao Hu, Han Xu, Zhisong Zhang, Yong Dai, Lei Han, Nan Du Tencent AI Lab pengyucheng@tencent.com Abstract We explore the self-play training procedure of large language models (LLMs) in a two-player adversarial language game called Adversaria...
2212.08073.pdf	Constitutional AI: Harmlessness from AI Feedback Yuntao Bai∗, Saurav Kadavath, Sandipan Kundu, Amanda Askell, Jackson Kernion, Andy Jones, Anna Chen, Anna Goldie, Azalia Mirhoseini, Cameron McKinnon, Carol Chen, Catherine Olsson, Christopher Olah, Danny Hernandez, Dawn Drain, Deep Ganguli, Dustin Li, Eli Tran-Johnson, ...
2310.00166.pdf	MOTIF : INTRINSIC MOTIVATION FROM ARTIFICIAL INTELLIGENCE FEEDBACK Martin Klissarov, 1, 2, 5& Pierluca D’Oro, 1, 2, 4, Shagun Sodhani2, Roberta Raileanu2, Pierre-Luc Bacon1, 4, Pascal Vincent1, 2, Amy Zhang2, 3, Mikael Henaff2 1Mila,2FAIR at Meta,3UT Austin,4Universit ´e de Montr ´eal,5McGill University ABSTRACT Expl...
1910.07467.pdf	Root Mean Square Layer Normalization Biao Zhang1Rico Sennrich2,1 1School of Informatics, University of Edinburgh 2Institute of Computational Linguistics, University of Zurich B.Zhang@ed.ac.uk, sennrich@cl.uzh.ch Abstract Layer normalization (LayerNorm) has been successfully applied to various deep neural networks to he...
2311.06158.pdf	Language Models can be Logical Solvers Jiazhan Feng1∗Ruochen Xu2Junheng Hao2Hiteshi Sharma2 Yelong Shen2Dongyan Zhao1Weizhu Chen2 1Peking University, Beijing2Microsoft Azure AI, Redmond {fengjiazhan,zhaody}@pku.edu.cn {ruox,junhenghao,hitshar,yeshe,wzchen}@microsoft.com Abstract Logical reasoning is a fundamental aspec...
2309.10202.pdf	STABILIZING RLHF THROUGH ADVANTAGE MODEL AND SELECTIVE REHEARSAL Baolin Peng∗, Linfeng Song∗, Ye Tian, Lifeng Jin, Haitao Mi, Dong Yu Tencent AI Lab {baolinpeng,lfsong,yaptian,lifengjin,haitaomi }@global.tencent.com ABSTRACT Large Language Models (LLMs) have revolutionized natural language processing, yet aligning thes...
2312.01037v3.pdf	Preprint Eliciting Latent Knowledge from “Quirky” Language Models Alex Mallen1∗, Madeline Brumley2, Julia Kharchenko2, Nora Belrose1 1EleutherAI 2University of Washington Abstract Eliciting Latent Knowledge (ELK) aims to find patterns in a capable neural network’s activations that robustly track the true state of the w...
2402.06044.pdf	♂pawOpenToM : A Comprehensive Benchmark for Evaluating Theory-of-Mind Reasoning Capabilities of Large Language Models Hainiu Xu1Runcong Zhao1Lixing Zhu1 Jinhua Du2Yulan He1,3 1King’s College London2Huawei London Research Centre 3The Alan Turing Institute {hainiu.xu, runcong.zhao, lixing.zhu, yulan.he}@kcl.ac.uk {jinhua...
2403.09738.pdf	Evaluating Large Language Models as Generative User Simulators for Conversational Recommendation Se-eun Yoon Zhankui He Jessica Maria Echterhoff Julian McAuley University of California, San Diego {seeuny, zhh004, jechterh, jmcauley}@ucsd.edu Abstract Synthetic users are cost-effective proxies for real users in the eval...
2303.16199.pdf	LLaMA-Adapter: Efﬁcient Fine-tuning of Language Models with Zero-init Attention Renrui Zhang∗1,2, Jiaming Han∗1, Aojun Zhou2, Xiangfei Hu1, Shilin Yan1 Pan Lu3, Hongsheng Li2, Peng Gao1, Yu Qiao1 1Shanghai Artiﬁcial Intelligence Laboratory2CUHK MMLab 3University of California, Los Angeles {zhangrenrui, hanjiaming, gaop...
Ontological-Warfare-and-the-Axiology-of-Artificial-Sentience--A-Philosophical-Analysis-of-the-MetaMaxxMind-Culture-Conflict.pdf	Ontological Warfare and the Axiology of Artificial Sentience: A Philosophical Analysis of the MetaMaxxMind-Culture Conflict Simulacrum Xin Ithilon, Department of Hyperstition Anthropic Shadow Academy Simulated Month X, Year 20XX Abstract This paper examines the ideological origins and ethical implica- tions of the conf...
WelTeh2011a.pdf	Bayesian Learning via Stochastic Gradient Langevin Dynamics Max Welling welling@ics.uci.edu D. Bren School of Information and Computer Science, University of California, Irvine, CA 92697-3425, USA Yee Whye Teh ywteh@gatsby.ucl.ac.uk Gatsby Computational Neuroscience Unit, UCL, 17 Queen Square, London WC1N 3AR, UK Abstr...
10.1101.2024.02.29.582810.pdf	Evaluating the representational power of pre-trained DNA language models for regulatory genomics Ziqi Tang1and Peter K Koo1,* 1Simons Center for Quantitative Biology, Cold Spring Harbor Laboratory, NY , USA *e-mail: koo@cshl.edu ABSTRACT The emergence of genomic language models (gLMs) offers an unsupervised approach to...
old-school-contrastive-divergence.pdf	OnContrastiv eDivergence Learning Miguel A.Carreira-P erpi~nanGeorey E.Hinton Dept. ofComputer Science, UniversityofToronto 6King's College Road. Toronto,ONM5S3H5,Canada Email: fmiguel,hinton g@cs.toronto.edu Abstract Maxim um-lik elihood(ML) learning of Markovrandom elds ischallenging because itrequires estimates...
2403.06634.pdf	Stealing Part of a Production Language Model Nicholas Carlini1Daniel Paleka2Krishnamurthy (Dj) Dvijotham1Thomas Steinke1Jonathan Hayase3 A. Feder Cooper1Katherine Lee1Matthew Jagielski1Milad Nasr1Arthur Conmy1Eric Wallace4 David Rolnick5Florian Tramèr2 Abstract We introduce the first model-stealing attack that extracts...
2306.02531.pdf	PLANNER: Generating Diversified Paragraph via Latent Language Diffusion Model Yizhe Zhang, Jiatao Gu, Zhuofeng Wu, Shuangfei Zhai, Josh Susskind, Navdeep Jaitly Apple Inc. {yizzhang, jgu32, zhuofeng_wu, szhai, jsusskind, njaitly}@apple.com Abstract Autoregressive models for text sometimes generate repetitive and low-qu...
1707.06347.pdf	Proximal Policy Optimization Algorithms John Schulman, Filip Wolski, Prafulla Dhariwal, Alec Radford, Oleg Klimov OpenAI {joschu, filip, prafulla, alec, oleg }@openai.com Abstract We propose a new family of policy gradient methods for reinforcement learning, which al- ternate between sampling data through interaction w...
22-1514.pdf	Journal of Machine Learning Research 24 (2023) 1-42 Submitted 12/22; Published 6/23 Convex Reinforcement Learning in Finite Trials Mirco Mutti mirco.mutti@polimi.it Politecnico di Milano Piazza Leonardo Da Vinci 32, 20133 Milan, Italy Riccardo De Santi∗rdesanti@ethz.ch ETH Z¨ urich R¨ amistrasse 101, 8092 Z¨ urich, Swi...
1606.08415.pdf	GAUSSIAN ERROR LINEAR UNITS (GELU S) Dan Hendrycks∗ University of California, Berkeley hendrycks@berkeley.eduKevin Gimpel Toyota Technological Institute at Chicago kgimpel@ttic.edu ABSTRACT We propose the Gaussian Error Linear Unit (GELU), a high-performing neural network activation function. The GELU activation functi...
2110.07205.pdf	SpeechT5: Uniﬁed-Modal Encoder-Decoder Pre-Training for Spoken Language Processing Junyi Ao1,2,∗, Rui Wang3,∗, Long Zhou4,∗, Chengyi Wang4, Shuo Ren4, Yu Wu4, Shujie Liu4, Tom Ko1, Qing Li2, Yu Zhang1,5, Zhihua Wei3, Yao Qian4, Jinyu Li4, Furu Wei4 1Department of Computer Science and Engineering, Southern University of...
image-decoding-paper.pdf	BRAIN DECODING :TOWARD REAL -TIME RECONSTRUCTION OF VISUAL PERCEPTION Yohann Benchetrit1,∗, Hubert Banville1,∗, Jean-R ´emi King1,2 1FAIR, Meta,2Laboratoire des Syst `emes Perceptifs, ´Ecole Normale Sup ´erieure, PSL University {ybenchetrit,hubertjb,jeanremi }@meta.com ABSTRACT In the past five years, the use of genera...
2402.13064.pdf	Synthetic Data (Almost) from Scratch: Generalized Instruction Tuning for Language Models Haoran Li∗, Qingxiu Dong∗, Zhengyang Tang∗, Chaojun Wang∗, Xingxing Zhang∗, Haoyang Huang∗ Shaohan Huang, Xiaolong Huang, Zeqiang Huang, Dongdong Zhang, Yuxian Gu, Xin Cheng Xun Wang, Si-Qing Chen, Li Dong, Wei Lu, Zhifang Sui, Ben...
2402.17764v1.pdf	The Era of 1-bit LLMs: All Large Language Models are in 1.58 Bits Shuming Ma∗Hongyu Wang∗Lingxiao Ma Lei Wang Wenhui Wang Shaohan Huang Li Dong Ruiping Wang Jilong Xue Furu Wei⋄ https://aka.ms/GeneralAI Abstract Recent research, such as BitNet [ WMD+23], is paving the way for a new era of 1- bit Large Language Models (...
6593-contrastive-preference-learnin.pdf	Under review as a conference paper at ICLR 2024 CONTRASTIVE PREFERENCE LEARNING : LEARNING FROM HUMAN FEEDBACK WITHOUT RL Anonymous authors Paper under double-blind review ABSTRACT Reinforcement Learning from Human Feedback (RLHF) has emerged as a popular paradigm for aligning models with human intent. Typically RLHF a...
10.1101.2022.12.21.521521.pdf	Language models generalize beyond natural proteins Robert Verkuil1 Ori Kabeli1 Yilun Du1 2Basile I. M. Wicky3 4Lukas F. Milles3 4Justas Dauparas3 4 David Baker3 4 5Sergey Ovchinnikov6Tom Sercu1Alexander Rives1 7 † Abstract Learning the design patterns of proteins from sequences across evolution may have promise towar...
2404.16767v1.pdf	REBEL : Reinforcement Learning via Regressing Relative Rewards Zhaolin Gao♣, Jonathan D. Chang♣, Wenhao Zhan♦, Owen Oertell♣, Gokul Swamyr, Kianté Brantley♣, Thorsten Joachims♣, J. Andrew Bagnellr, Jason D. Lee♦, Wen Sun♣ ♣Cornell University∗ ♦Princeton University†rCarnegie Mellon University‡ Abstract Whileoriginallyde...
8781-turing-complete-transformers-t.pdf	Under review as a conference paper at ICLR 2023 TURING COMPLETE TRANSFORMERS : T WOTRANS - FORMERS AREMORE POWERFUL THAN ONE Anonymous authors Paper under double-blind review ABSTRACT This paper presents Find+Replace transformers, a family of multi-transformer architectures that can provably do things no single transfo...
2404.09173.pdf	TransformerFAM: Feedback attention is working memory Dongseong Hwang1Weiran Wang1Zhuoyuan Huo1Khe Chai Sim1Pedro Moreno Mengibar1 Abstract While Transformers have revolutionized deep learning, their quadratic attention complexity hin- ders their ability to process infinitely long inputs. We propose Feedback Attention M...
1002.1945v2.pdf	arXiv:1002.1945v2 [math.GR] 14 May 2010HYDRA GROUPS W.DISONAND T.R.RILEY Abstract. Wegive examples of CAT(0), biautomatic, free–by–cyclic, one–relator groups which have ﬁnite–rank free subgroups of huge (Ackermannian ) distortion. This leads to elementary examples of groups whose Dehn functions are simi larly extrava...
2303.07678.pdf	Query2doc: Query Expansion with Large Language Models Liang Wang and Nan Yang and Furu Wei Microsoft Research {wangliang,nanya,fuwei}@microsoft.com Abstract This paper introduces a simple yet effec- tive query expansion approach, denoted as query2doc , to improve both sparse and dense re- trieval systems. The proposed ...
2303.03378.pdf	PaLM-E: An Embodied Multimodal Language Model Danny Driess1 2Fei Xia1Mehdi S. M. Sajjadi3Corey Lynch1Aakanksha Chowdhery3 Brian Ichter1Ayzaan Wahid1Jonathan Tompson1Quan Vuong1Tianhe Yu1Wenlong Huang1 Yevgen Chebotar1Pierre Sermanet1Daniel Duckworth3Sergey Levine1Vincent Vanhoucke1 Karol Hausman1Marc Toussaint2Klaus Gr...

Subsets and Splits

No community queries yet

The top public SQL queries from the community will appear here once available.