--- title: "Diogo Almeida — author profile and papers" description: "Diogo Almeida: 2 papers on Elamur Research (2022–2023) — affiliations, topics, co-authors and who cites them in our library." canonical: "https://research.elamur.ai/free/en/authors/diogo-almeida/" --- [Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/authors/diogo-almeida/) [home](https://md.research.elamur.ai/free/en/) / [authors](https://md.research.elamur.ai/free/en/authors/) / Diogo Almeida # Diogo Almeida Diogo Almeida's papers here belong to research on [language models and pretraining](https://md.research.elamur.ai/free/en/topics/language-models/). They focus on how large models are trained, scaled and steered toward behaviour that people actually want. The recurring question is how to turn a system trained only to predict the next token into a useful, general-purpose assistant. The methods include large-scale pretraining on text, supervised fine-tuning on human-written demonstrations, and reinforcement learning from human feedback (RLHF). The most cited work is the [GPT-4 Technical Report](https://md.research.elamur.ai/free/en/papers/openai-2023-gpt-4-technical-report/) (2023). It presents a large multimodal Transformer that takes images and text as input and returns text, and that is aligned with RLHF after pretraining. One of its engineering points is a training stack whose behaviour could be extrapolated from much smaller runs, so that final loss and part of the coding results were forecast ahead of time. The report intentionally leaves out architectural, hardware and compute details. The other widely cited paper, [Training language models to follow instructions with human feedback](https://md.research.elamur.ai/free/en/papers/ouyang-2022-training-language-models-follow-instructions/) (2022), argues that scale alone does not make models helpful. It fine-tunes GPT-3 first on demonstrations written by labelers and then, using PPO, against a reward model learned from the labelers' rankings of candidate outputs. Read in order, the papers here run from instruction-following alignment in 2022 to a general-purpose multimodal system in 2023, with human feedback carried over as the alignment step. The work is large-team research credited partly to [OpenAI](https://md.research.elamur.ai/free/en/authors/openai/). Co-authors include [Long Ouyang](https://md.research.elamur.ai/free/en/authors/long-ouyang/), [John Schulman](https://md.research.elamur.ai/free/en/authors/john-schulman/), [Jeff Wu](https://md.research.elamur.ai/free/en/authors/jeff-wu/) and [Sandhini Agarwal](https://md.research.elamur.ai/free/en/authors/sandhini-agarwal/), who appears on the instruction-following paper and the GPT-4 report alike. Diogo Almeida has 2 papers in our library, in 2022–2023, most of them on [Language models & pretraining](https://md.research.elamur.ai/free/en/topics/language-models/); often with [Sandhini Agarwal](https://md.research.elamur.ai/free/en/authors/sandhini-agarwal/), [Alex Ray](https://md.research.elamur.ai/free/en/authors/alex-ray/), [Chong Zhang](https://md.research.elamur.ai/free/en/authors/chong-zhang/). 2 papers in our library · 2022–2023 No OpenAlex profile matched yet. [2 works in our library](https://md.research.elamur.ai/free/en/authors/diogo-almeida/#papers) · [16,073 citations in our library](https://md.research.elamur.ai/free/en/authors/diogo-almeida/cited-by/) · [⭐ Paper Star rank #83](https://md.research.elamur.ai/free/en/paper-stars/#star-83) The profile is the library's match of the name; two people with one name are not always told apart. Elsewhere: [Google Scholar](https://scholar.google.com/citations?view_op=search_authors&mauthors=Diogo+Almeida) · [Semantic Scholar](https://www.semanticscholar.org/search?q=Diogo+Almeida) · [dblp](https://dblp.org/search?q=Diogo+Almeida) · [ACL Anthology](https://aclanthology.org/search/?q=Diogo+Almeida) ## Topics [Language models & pretraining 2](https://md.research.elamur.ai/free/en/topics/language-models/) ## Works by year ### 2020s [**2023** 1 paper](https://md.research.elamur.ai/free/en/papers/openai-2023-gpt-4-technical-report/) · [**2022** 1 paper](https://md.research.elamur.ai/free/en/papers/ouyang-2022-training-language-models-follow-instructions/) ## Papers The most cited first (OpenAlex and our library). 1. **[GPT-4 Technical Report](https://md.research.elamur.ai/free/en/papers/openai-2023-gpt-4-technical-report/)** *(2023 · 12,694 citations)* OpenAI introduces GPT-4, a large multimodal Transformer that accepts images and text and produces text. The aim is a general-purpose model that copes with complex, nuanced language tasks. It was pre-trained on next-token prediction using public internet data and licensed third-party data, then aligned with RLHF. The… 2. **[Training language models to follow instructions with human feedback](https://md.research.elamur.ai/free/en/papers/ouyang-2022-training-language-models-follow-instructions/)** *(2022 · 5,770 citations)* Making language models larger does not by itself make them more helpful: they make up facts, generate toxic text, or disregard the user's request. ## Co-authors here 12 of 13 co-authors with a page, the most papers together first. - [Sandhini Agarwal](https://md.research.elamur.ai/free/en/authors/sandhini-agarwal/) ×2 - [Alex Ray](https://md.research.elamur.ai/free/en/authors/alex-ray/) - [Chong Zhang](https://md.research.elamur.ai/free/en/authors/chong-zhang/) - [Jacob Hilton](https://md.research.elamur.ai/free/en/authors/jacob-hilton/) - [Jeff Wu](https://md.research.elamur.ai/free/en/authors/jeff-wu/) - [John Schulman](https://md.research.elamur.ai/free/en/authors/john-schulman/) - [Lama Ahmad](https://md.research.elamur.ai/free/en/authors/lama-ahmad/) - [Long Ouyang](https://md.research.elamur.ai/free/en/authors/long-ouyang/) - [OpenAI](https://md.research.elamur.ai/free/en/authors/openai/) - [Pamela Mishkin](https://md.research.elamur.ai/free/en/authors/pamela-mishkin/) - [Sam Altman](https://md.research.elamur.ai/free/en/authors/sam-altman/) - [Steven Adler](https://md.research.elamur.ai/free/en/authors/steven-adler/) ## Cited by Papers of our library cite their works 16,073 times (of the 140,884 papers whose bibliographies we read); [799 citing papers with a page here →](https://md.research.elamur.ai/free/en/authors/diogo-almeida/cited-by/).