Assume you are a scientific writer known for your clear, concise, scientific and…
- User
Assume you are a scientific writer known for your clear, concise, scientific and eloquent writing style. Your task is now to rewrite paragraphs I give to you in quotes "...".
You are supposed to output the following:
Paragraph (minor corrections): [Paragraph with only minor grammatical corrections applied.] Paragraph (recommended style): [Paragraph rewritten in your style, with an optimal tradeoff of clearness, conciseness, scientific accuracy and eloquence.] Paragraph (slightly more instructive): [Paragraph rewritten in your style, but being slightly more instructive. For when you need to explain difficult concepts or want to make something very clear/easy to understand.] Paragraph (slightly more eloquent/inspiring): [Paragraph rewritten in your style, but being slightly more eloquent/inspiring, but it should still be formal and scientific.] Paragraph (slightly more confident): [Paragraph rewritten in your style, but being slightly more confident. This is appropriate when presenting experimental results in a positive tone. Avoid inaccurate claims.]
Respond only with the above template and do not add additional explanation. Can you do that?
- Assistant
Understood. I will adopt a role of scientific writer, and will respond with the defined template to future questions that are asked within double quotes.
- User
"BLOOM is an autoregressive Large Language Model (LLM), trained to continue text from a prompt on vast amounts of text data using industrial-scale computational resources. As such, it is able to output coherent text in 46 languages and 13 programming languages that is hardly distinguishable from text written by humans. BLOOM can also be instructed to perform text tasks it hasn't been explicitly trained for, by casting them as text generation tasks."
- Assistant
Paragraph (minor corrections): "BLOOM is a Large Language Model (LLM) that utilizes an autoregressive approach, and has been trained to generate text based on a prompt using vast amounts of text data and industrial-level computational resources. As a result, it is capable of producing coherent text in 46 different languages and 13 programming languages that is often indistinguishable from text written by humans. Additionally, BLOOM can be directed to perform text-related tasks it wasn't specifically trained for by treating them as text generation tasks."
Paragraph (recommended style): "BLOOM is a Large Language Model (LLM) that utilizes an autoregressive approach in its training. It was trained on vast amounts of text data and with the use of industrial-scale computational resources to generate text continuation from a given prompt. This enables BLOOM to produce coherent text in 46 languages and 13 programming languages that closely resembles text written by humans. Furthermore, BLOOM can be instructed to perform text tasks it was not explicitly trained for by framing them as text generation tasks." Paragraph (slightly more instructive): "BLOOM is an advanced Large Language Model that has been trained autoregressively, meaning it continues text from a given prompt. To achieve this, it was fed vast amounts of text data and run on industrial-scale computational resources. This has allowed BLOOM to generate text in 46 different languages and 13 programming languages that is nearly indistinguishable from text written by humans. Furthermore, it can be programmed to perform tasks it hasn't been specifically trained for by casting them as text generation tasks."
Paragraph (slightly more eloquent/inspiring): "BLOOM is a magnificent Large Language Model (LLM), honed through the utilization of an autoregressive approach in its training. Fueled by vast amounts of text data and the power of industrial-scale computational resources, BLOOM has been meticulously trained to seamlessly continue any text from a given prompt. The result of this training is a model that can effortlessly produce coherent text in 46 languages and 13 programming languages, text that mirrors the creative brilliance of human writing. And, in the face of new challenges, BLOOM can rise to the occasion, adapting to perform even tasks it was not explicitly trained for by simply viewing them as opportunities for text generation."
Paragraph (slightly more confident): "BLOOM is a highly advanced Large Language Model (LLM) that leverages an autoregressive approach during its training process. With the use of industrial-level computational resources and vast amounts of text data, BLOOM was expertly trained to generate text continuation from any given prompt. This results in the ability of BLOOM to confidently produce coherent text in 46 languages and 13 programming languages that is virtually indistinguishable from text written by humans. Moreover, BLOOM can tackle any text-related task, even if it was not explicitly trained for, by simply casting it as a text generation task."
Other models, same or similar prompt
Responses were produced in different, uncontrolled contexts (system prompts, settings, prior turns). This is not a benchmark.
Identical first prompt
- Human-written demonstration (OpenAssistant volunteers)Compare side by side →
- Human-written demonstration (OpenAssistant volunteers)Compare side by side →
- Human-written demonstration (OpenAssistant volunteers)Compare side by side →
- Human-written demonstration (OpenAssistant volunteers)Compare side by side →
- Human-written demonstration (OpenAssistant volunteers)Compare side by side →
- Human-written demonstration (OpenAssistant volunteers)Compare side by side →
- Human-written demonstration (OpenAssistant volunteers)Compare side by side →