Ricardo Primi

Psychologist data scientist with a focus on psychometrics. Professor at USF, Campinas. Visiting scholar at UC Berkeley and Lemann Center at Stanford University

[Ricardo Primi @rprimi]

Hawkins vs. The Upside Down: The Psychology Edition. 🚲🔴

[Ricardo Primi @rprimi]

@MelMitchell1 I'm really excited about the upcoming new content. Can't wait to start reading it!

[Ricardo Primi @rprimi]

Evento da Lemann trouxe Sam Altman para o Rio. Vamos aproveitar essa onda revolucionaria que a IA esta trazendo e aplicar na melhora da educação no Brasil.

[Ricardo Primi @rprimi]

See the report here https://t.co/uPaZCqHdQV /04/chatgpt-e-melhor-que-80-dos-alunos-no-enem-mas-derrapa-em-matematica.shtml?utm_source=twitter&utm_medium=social&utm_campaign=comptw

[Ricardo Primi @rprimi]

Our paper with @desnes, @piresramon, @rodrigfnogueira, @robertoalotufo demonstrating the performance of GPT 3.5 and GPT-4 on ENEM questions (score of 87%) was quoted in today's folha de são paulo in the report from @_danielmariani see the paper here
https://t.co/xM3oWIdlOV

[Ricardo Primi @rprimi]

An interesting result is that GPT-4 with examples of "Chain-of-Thought" did better than GPT 3.5 in questions of mathematical reasoning See more details here https://t.co/6AQYUsPdaD

[Rodrigo Nogueira @rodrigfnogueira]

https://t.co/M0zTTkP7fS
We evaluated GPT-4 on a widely used admission exam to Brazilian Universities, and the results were very good: it answered correctly 80% of the questions that do not require image interpretation (the GPT-4 version that supports images is not yet available).

[Rodrigo Nogueira @rodrigfnogueira]

https://t.co/M0zTTkP7fS
We evaluated GPT-4 on a widely used admission exam to Brazilian Universities, and the results were very good: it answered correctly 80% of the questions that do not require image interpretation (the GPT-4 version that supports images is not yet available).

[Ricardo Primi @rprimi]

@felipeneto Vai ver que tava acabando a poção de polissuco ..

[Beatriz Milz @BeaMilz]

Thankfully Julio asked for a picture, otherwise no one would believe that we had lunch with @hadleywickham and @chrisderv 😄
@jtrecenti @flaviaerius @rprimi e @anakiki1505 ! #rstudioconf2022 #RStudioConf #posit @rstudio #RStats

[Ricardo Primi @rprimi]

Wow!! Shiny for python !!

[Ricardo Primi @rprimi]

At amazing #rstudioconf2022

[Leaflet @LeafletJS]

I'm a little mapping library that was born 11 years ago in Ukraine. 🇺🇦 Yesterday, after enduring 8 years of hybrid warfare that killed 15000+ and displaced millions, Ukrainians listened to a 40-minute hate speech from Russia's insane president on why Ukraine shouldn't exist. 1/4

[Paulo Saldaña @paulosaldana]

A ideia dessa pauta veio há meses e foi possível graças ao Daniel Mariani, do @DeltaFolha, q fez mergulho na TRI como poucos já fizeram. Agradecimento especial aos prof. Tufi Machado e @rprimi. A primeira matéria desse esforço foi esta abaixo. E vem mais
https://t.co/MvjFGMElUm

[Ricardo Primi @rprimi]

Candidato do Enem, comece sempre pelas perguntas fáceis https://t.co/5GoBy9p3Li

[Professora Dayane Pimentel @deppimentel]

O 1:27 mais completos que eu já ouvi. Senador @Sen_Alessandro resumindo a farsa Bolsonaro.

[Ricardo Primi @rprimi]

@mcxfrank You need to create an acquiescence index per subject, use it to remove item variance due to acquiescence, and run item EFA again with these residualized item scores. aWe explain this method here https://t.co/e7uCCczH21

[Ricardo Primi @rprimi]

@mcxfrank Yes, these items help. These negative items factor emerges from suppression/inflation effects of acquiescence on item scores. See this: https://t.co/QNmXbd3ooD

[Ricardo Primi @rprimi]

#parlamentares_levantem_da_mesa

[Tavinhu Furtado @tavinhu_furtado]

Todo meu respeito por @TaschnerNatalia, né m especial pela coragem dessa fala 👏🏻

[Tavinhu Furtado @tavinhu_furtado]

Todo meu respeito por @TaschnerNatalia, né m especial pela coragem dessa fala 👏🏻