Meta tags:
description=Jeffrey Cheng;

Headings (most frequently used words):

-

Text of the page (most frequently used words):
and (11), models (8), reasoning (7), data (6), txt (6), pretraining (5), about (5), phd (4), princeton (4), the (4), how (4), language (4), 2024 (3), research (3), advised (3), can (3), rewards (3), pli (3), cheng (3), dated (2), compressed (2), chain (2), thought (2), university (2), interests (2), with (2), make (2), more (2), efficient (2), from (2), space (2), knowledge (2), interested (2), contact (2), jeffrey (2), scholar (2), publications (2), attended, colm, presented, wins, outstanding, paper, award, top, oct, new, preprint, released, dec, started, graciously, supported, francis, upton, fellowship, aug, 2025, news, outside, avid, player, also, starting, run, again, after, long, break, chess, climber, misc, previously, recieved, master, johns, hopkins, prior, nlp, were, mathematics, fluid, dynamics, conducted, these, areas, during, undergraduate, studies, duke, tarek, elgindi, benjamin, van, durme, previous, construct, environments, verifiable, induce, structure, into, chains, capable, much, better, would, trained, process, rather, than, just, outcome, shifting, away, discrete, token, perform, continuous, latent, best, correct, misalignments, arising, conflicts, attribute, content, generated, back, their, corpus, does, influence, sources, below, are, few, questions, first, year, student, intelligence, broadly, lie, intersection, natural, processing, machine, learning, currently, agents, particular, aim, study, downstream, effects, methods, improve, capabilities, efficiency, danqi, chen, via, email, jc93, dot, edu, orcid, linkedin, twitter, github, follow, photo, jpg, jeff,

Text of the page (random words):
about me jeffrey cheng jeff cheng about me publications cv about me publications cv photo jpg contact txt phd pli follow github twitter g scholar linkedin s scholar orcid jeffrey cheng phd princeton pli contact me via x or email jc93 at princeton dot edu about me about txt i am a first year phd student at princeton language and intelligence pli advised by danqi chen my research interests broadly lie in the intersection of natural language processing and machine learning i am currently interested in language models and agents in particular i aim to study the downstream effects of pretraining data and methods to improve the capabilities and efficiency of reasoning models below are a few questions i am interested in pretraining txt data how does pretraining data influence language models as sources of knowledge dated data can we attribute content generated by models back to their pretraining corpus how do we best correct misalignments arising from knowledge conflicts in models pretraining data reasoning txt reasoning can we make reasoning models more efficient by shifting away from a discrete token space and perform reasoning in continuous latent space compressed chain of thought how much better would reasoning models be if trained with process rewards rather than just outcome rewards how can we construct environments with verifiable rewards and or induce structure into reasoning chains to make models more capable and efficient previous txt previously i recieved my master s at johns hopkins university advised by benjamin van durme prior to nlp my interests were in mathematics and fluid dynamics i conducted research in these areas during my undergraduate studies at duke university advised by tarek elgindi misc txt outside of research i am an avid climber and chess player i am also starting to run again after a long break news aug 2025 started my phd at princeton graciously supported by the francis upton fellowship dec 2024 new preprint compressed chain of thought released oct 2024 attended colm 2024 and presented dated data it wins outstanding paper award top 0 4