# RLHF jobs online (12 figures) > 6 of 12 platforms whose own pages we read describe ranking, comparing or rating AI model responses, the core of RLHF work, read 22 October 2026. RLHFJobs · https://rlhfjobs.com/datasets/rlhf-task-types/ · computed 2026-10-22 **Describe ranking, comparing or rating responses:** 6 of 12 platforms (15 pages read; 22 named). ## The answers, question by question - **Ranks, compares or rates model responses:** not stated 6, describes ranking, comparing or rating responses 6 - **Writes prompts, answers or rubrics:** not stated 9, describes writing prompts, answers or rubrics, or rewriting responses 3 - **Red-teams models:** not stated 11, describes red-teaming or probing models for weaknesses 1 - **Domain tracks it names, in its own words:** stated 6, not stated 6 - **Says its general work needs no experience:** not stated 5, says some work needs no experience or special skills 4, asks for domain expertise, a degree or professional experience 2, varies 1 - **Asks an entry assessment:** not stated 7, asks an entry assessment, screening or qualification task 5 ## Every platform | Platform | Ranks, compares or rates model responses | Writes prompts, answers or rubrics | Red-teams models | Domain tracks it names, in its own words | Says its general work needs no experience | Asks an entry assessment | |---|---|---|---|---|---|---| | Alignerr | not stated | not stated | describes red-teaming or probing models for weaknesses | View all roles Audio Coding General STEM | says some work needs no experience or special skills | not stated | | Clickworker | not stated | not stated | not stated | not stated | not stated | not stated | | CrowdGen | describes ranking, comparing or rating responses | not stated | not stated | not stated | asks for domain expertise, a degree or professional experience | asks an entry assessment, screening or qualification task | | DataAnnotation | describes ranking, comparing or rating responses | describes writing prompts, answers or rubrics, or rewriting responses | not stated | Coding Generalist Law Mathematics Medicine Physics Finance Accounting Bilingual Chemistry… | says some work needs no experience or special skills | asks an entry assessment, screening or qualification task | | Handshake AI | not stated | not stated | not stated | not stated | says some work needs no experience or special skills | not stated | | Mercor | not stated | not stated | not stated | not stated | not stated | asks an entry assessment, screening or qualification task | | Mindrift | describes ranking, comparing or rating responses | describes writing prompts, answers or rubrics, or rewriting responses | not stated | We accept experts across 90+ domains from STEM and law to creative writing and… | varies by document | asks an entry assessment, screening or qualification task | | OneForma | describes ranking, comparing or rating responses | not stated | not stated | Marketing, psychology, journalism, architecture, generalist. If you know it, we need it. | not stated | not stated | | Outlier | describes ranking, comparing or rating responses | describes writing prompts, answers or rubrics, or rewriting responses | not stated | We depend on subject matter experts in coding, STEM, languages, and more to improve AI | asks for domain expertise, a degree or professional experience | asks an entry assessment, screening or qualification task | | Prolific | describes ranking, comparing or rating responses | not stated | not stated | not stated | not stated | not stated | | RemoExperts | not stated | not stated | not stated | Cinematographers Photoshop Experts After Effects Specialists Coding Developers STEM PhDs… | not stated | not stated | | Remotasks | not stated | not stated | not stated | not stated | says some work needs no experience or special skills | not stated | ## The sentence each answer rests on (90) - Alignerr, [Home page](https://www.alignerr.com/), read 2026-10-22, Ranks, compares or rates model responses: Nothing on this page answers it (1,876 characters of page text, read 2026-10-22). - Alignerr, [Home page](https://www.alignerr.com/), read 2026-10-22, Writes prompts, answers or rubrics: Nothing on this page answers it (1,876 characters of page text, read 2026-10-22). - Alignerr, [Home page](https://www.alignerr.com/), read 2026-10-22, Red-teams models: "Probe cutting-edge models for weaknesses and help make them safer" - Alignerr, [Home page](https://www.alignerr.com/), read 2026-10-22, Domain tracks it names, in its own words: "View all roles Audio Coding General STEM" - Alignerr, [Home page](https://www.alignerr.com/), read 2026-10-22, Says its general work needs no experience: "No tech background needed, just a camera and curiosity" - Alignerr, [Home page](https://www.alignerr.com/), read 2026-10-22, Asks an entry assessment: Nothing on this page answers it (1,876 characters of page text, read 2026-10-22). - Clickworker, [Clickworker page](https://www.clickworker.com/clickworker/), read 2026-10-08, Ranks, compares or rates model responses: Nothing on this page answers it (9,822 characters of page text, read 2026-10-08). - Clickworker, [Clickworker page](https://www.clickworker.com/clickworker/), read 2026-10-08, Writes prompts, answers or rubrics: Nothing on this page answers it (9,822 characters of page text, read 2026-10-08). - Clickworker, [Clickworker page](https://www.clickworker.com/clickworker/), read 2026-10-08, Red-teams models: Nothing on this page answers it (9,822 characters of page text, read 2026-10-08). - Clickworker, [Clickworker page](https://www.clickworker.com/clickworker/), read 2026-10-08, Domain tracks it names, in its own words: Nothing on this page answers it (9,822 characters of page text, read 2026-10-08). - Clickworker, [Clickworker page](https://www.clickworker.com/clickworker/), read 2026-10-08, Says its general work needs no experience: Nothing on this page answers it (9,822 characters of page text, read 2026-10-08). - Clickworker, [Clickworker page](https://www.clickworker.com/clickworker/), read 2026-10-08, Asks an entry assessment: Nothing on this page answers it (9,822 characters of page text, read 2026-10-08). - CrowdGen, [Home page](https://crowdgen.com/), read 2026-10-08, Ranks, compares or rates model responses: "Evaluate, annotate, generate, and refine" - CrowdGen, [Home page](https://crowdgen.com/), read 2026-10-08, Writes prompts, answers or rubrics: Nothing on this page answers it (9,294 characters of page text, read 2026-10-08). - CrowdGen, [Home page](https://crowdgen.com/), read 2026-10-08, Red-teams models: Nothing on this page answers it (9,294 characters of page text, read 2026-10-08). - CrowdGen, [Home page](https://crowdgen.com/), read 2026-10-08, Domain tracks it names, in its own words: Nothing on this page answers it (9,294 characters of page text, read 2026-10-08). - CrowdGen, [Home page](https://crowdgen.com/), read 2026-10-08, Says its general work needs no experience: "Most of our projects don’t require technical skills, just domain expertise." - CrowdGen, [Home page](https://crowdgen.com/), read 2026-10-08, Asks an entry assessment: "Most of our projects have a qualification process." - DataAnnotation, [Home page](https://www.dataannotation.tech/), read 2026-10-22, Ranks, compares or rates model responses: "commonly include reviewing AI-generated text, ranking responses, checking facts" - DataAnnotation, [Home page](https://www.dataannotation.tech/), read 2026-10-22, Writes prompts, answers or rubrics: "like reviewing responses, checking accuracy, refining prompts, and rating outputs" - DataAnnotation, [Home page](https://www.dataannotation.tech/), read 2026-10-22, Red-teams models: Nothing on this page answers it (7,040 characters of page text, read 2026-10-22). - DataAnnotation, [Home page](https://www.dataannotation.tech/), read 2026-10-22, Domain tracks it names, in its own words: Nothing on this page answers it (7,040 characters of page text, read 2026-10-22). - DataAnnotation, [Home page](https://www.dataannotation.tech/), read 2026-10-22, Says its general work needs no experience: "No coding or technical background needed." - DataAnnotation, [Home page](https://www.dataannotation.tech/), read 2026-10-22, Asks an entry assessment: Nothing on this page answers it (7,040 characters of page text, read 2026-10-22). - DataAnnotation, [FAQ](https://www.dataannotation.tech/faq), read 2026-10-08, Ranks, compares or rates model responses: "you’ll evaluate chatbot responses for accuracy, compare AI outputs, flag factual errors" - DataAnnotation, [FAQ](https://www.dataannotation.tech/faq), read 2026-10-08, Writes prompts, answers or rubrics: "test AI-generated images, write challenging prompts, or review code for errors" - DataAnnotation, [FAQ](https://www.dataannotation.tech/faq), read 2026-10-08, Red-teams models: Nothing on this page answers it (27,215 characters of page text, read 2026-10-08). - DataAnnotation, [FAQ](https://www.dataannotation.tech/faq), read 2026-10-08, Domain tracks it names, in its own words: "Coding Generalist Law Mathematics Medicine Physics Finance Accounting Bilingual Chemistry Biology" - DataAnnotation, [FAQ](https://www.dataannotation.tech/faq), read 2026-10-08, Says its general work needs no experience: "No coding experience is required for general projects" - DataAnnotation, [FAQ](https://www.dataannotation.tech/faq), read 2026-10-08, Asks an entry assessment: "You can only take the Starter Assessment once" - Handshake AI, [Programme page](https://joinhandshake.com/ai/), read 2026-10-22, Ranks, compares or rates model responses: Nothing on this page answers it (9,523 characters of page text, read 2026-10-22). - Handshake AI, [Programme page](https://joinhandshake.com/ai/), read 2026-10-22, Writes prompts, answers or rubrics: Nothing on this page answers it (9,523 characters of page text, read 2026-10-22). - Handshake AI, [Programme page](https://joinhandshake.com/ai/), read 2026-10-22, Red-teams models: Nothing on this page answers it (9,523 characters of page text, read 2026-10-22). - Handshake AI, [Programme page](https://joinhandshake.com/ai/), read 2026-10-22, Domain tracks it names, in its own words: Nothing on this page answers it (9,523 characters of page text, read 2026-10-22). - Handshake AI, [Programme page](https://joinhandshake.com/ai/), read 2026-10-22, Says its general work needs no experience: "Get paid hourly for high-impact work, no matter your background." - Handshake AI, [Programme page](https://joinhandshake.com/ai/), read 2026-10-22, Asks an entry assessment: Nothing on this page answers it (9,523 characters of page text, read 2026-10-22). - Mercor, [Open roles page](https://work.mercor.com/), read 2026-10-08, Ranks, compares or rates model responses: Nothing on this page answers it (3,011 characters of page text, read 2026-10-08). - Mercor, [Open roles page](https://work.mercor.com/), read 2026-10-08, Writes prompts, answers or rubrics: Nothing on this page answers it (3,011 characters of page text, read 2026-10-08). - Mercor, [Open roles page](https://work.mercor.com/), read 2026-10-08, Red-teams models: Nothing on this page answers it (3,011 characters of page text, read 2026-10-08). - Mercor, [Open roles page](https://work.mercor.com/), read 2026-10-08, Domain tracks it names, in its own words: Nothing on this page answers it (3,011 characters of page text, read 2026-10-08). - Mercor, [Open roles page](https://work.mercor.com/), read 2026-10-08, Says its general work needs no experience: Nothing on this page answers it (3,011 characters of page text, read 2026-10-08). - Mercor, [Open roles page](https://work.mercor.com/), read 2026-10-08, Asks an entry assessment: "Get Instant work offers 1 Resume 2 Location 3 Assessment" - Mindrift, [Home page](https://mindrift.ai/), read 2026-10-22, Ranks, compares or rates model responses: "Writing expert prompts, evaluating AI-generated responses, justifying your assessments" - Mindrift, [Home page](https://mindrift.ai/), read 2026-10-22, Writes prompts, answers or rubrics: "Create realistic prompts, questions, and scenarios, the kind a real user would actually ask AI." - Mindrift, [Home page](https://mindrift.ai/), read 2026-10-22, Red-teams models: Nothing on this page answers it (4,004 characters of page text, read 2026-10-22). - Mindrift, [Home page](https://mindrift.ai/), read 2026-10-22, Domain tracks it names, in its own words: "We accept experts across 90+ domains from STEM and law to creative writing and linguistics." - Mindrift, [Home page](https://mindrift.ai/), read 2026-10-22, Says its general work needs no experience: "You need domain expertise" - Mindrift, [Home page](https://mindrift.ai/), read 2026-10-22, Asks an entry assessment: "Complete a technical assessment to demonstrate your coding skills" - Mindrift, [Welcome page for Toloka members](https://mindrift.ai/welcome), read 2026-10-08, Ranks, compares or rates model responses: "Compare two options, pick the better one, or rate by criteria like accuracy and helpfulness." - Mindrift, [Welcome page for Toloka members](https://mindrift.ai/welcome), read 2026-10-08, Writes prompts, answers or rubrics: Nothing on this page answers it (5,321 characters of page text, read 2026-10-08). - Mindrift, [Welcome page for Toloka members](https://mindrift.ai/welcome), read 2026-10-08, Red-teams models: Nothing on this page answers it (5,321 characters of page text, read 2026-10-08). - Mindrift, [Welcome page for Toloka members](https://mindrift.ai/welcome), read 2026-10-08, Domain tracks it names, in its own words: Nothing on this page answers it (5,321 characters of page text, read 2026-10-08). - Mindrift, [Welcome page for Toloka members](https://mindrift.ai/welcome), read 2026-10-08, Says its general work needs no experience: "You can start on All-access tasks right away." - Mindrift, [Welcome page for Toloka members](https://mindrift.ai/welcome), read 2026-10-08, Asks an entry assessment: "take short qualification tasks, and unlock projects suited to your expertise" - OneForma, [Home page](https://www.oneforma.com/), read 2026-10-22, Ranks, compares or rates model responses: "Apply your expertise to evaluate AI, flag what it gets wrong, and help shape what comes next." - OneForma, [Home page](https://www.oneforma.com/), read 2026-10-22, Writes prompts, answers or rubrics: Nothing on this page answers it (7,764 characters of page text, read 2026-10-22). - OneForma, [Home page](https://www.oneforma.com/), read 2026-10-22, Red-teams models: Nothing on this page answers it (7,764 characters of page text, read 2026-10-22). - OneForma, [Home page](https://www.oneforma.com/), read 2026-10-22, Domain tracks it names, in its own words: "Marketing, psychology, journalism, architecture, generalist. If you know it, we need it." - OneForma, [Home page](https://www.oneforma.com/), read 2026-10-22, Says its general work needs no experience: Nothing on this page answers it (7,764 characters of page text, read 2026-10-22). - OneForma, [Home page](https://www.oneforma.com/), read 2026-10-22, Asks an entry assessment: Nothing on this page answers it (7,764 characters of page text, read 2026-10-22). - Outlier, [Home page](https://outlier.ai/), read 2026-10-08, Ranks, compares or rates model responses: "Write a Challenging Prompt Create Grading Rubrics Rate and Rank Answers" - Outlier, [Home page](https://outlier.ai/), read 2026-10-08, Writes prompts, answers or rubrics: "Think of a difficult question in your field of study" - Outlier, [Home page](https://outlier.ai/), read 2026-10-08, Red-teams models: Nothing on this page answers it (7,956 characters of page text, read 2026-10-08). - Outlier, [Home page](https://outlier.ai/), read 2026-10-08, Domain tracks it names, in its own words: "We depend on subject matter experts in coding, STEM, languages, and more to improve AI" - Outlier, [Home page](https://outlier.ai/), read 2026-10-08, Says its general work needs no experience: "minimum qualifications typically require undergraduate-level expertise" - Outlier, [Home page](https://outlier.ai/), read 2026-10-08, Asks an entry assessment: "completing skill screenings so we can match you with the right projects" - Outlier, [FAQ](https://outlier.ai/faq), read 2026-10-22, Ranks, compares or rates model responses: "you may generate prompts, rank AI outputs, improve model responses" - Outlier, [FAQ](https://outlier.ai/faq), read 2026-10-22, Writes prompts, answers or rubrics: "you may generate prompts, rank AI outputs, improve model responses" - Outlier, [FAQ](https://outlier.ai/faq), read 2026-10-22, Red-teams models: Nothing on this page answers it (6,004 characters of page text, read 2026-10-22). - Outlier, [FAQ](https://outlier.ai/faq), read 2026-10-22, Domain tracks it names, in its own words: Nothing on this page answers it (6,004 characters of page text, read 2026-10-22). - Outlier, [FAQ](https://outlier.ai/faq), read 2026-10-22, Says its general work needs no experience: "we seek experts with deep knowledge, advanced skills, and significant experience in their fields" - Outlier, [FAQ](https://outlier.ai/faq), read 2026-10-22, Asks an entry assessment: "Review project guidelines and complete assessment tasks" - Prolific, [Participants page](https://www.prolific.com/participants), read 2026-10-08, Ranks, compares or rates model responses: "Evaluate machine learning models and consider key questions about AI and human interaction" - Prolific, [Participants page](https://www.prolific.com/participants), read 2026-10-08, Writes prompts, answers or rubrics: Nothing on this page answers it (5,958 characters of page text, read 2026-10-08). - Prolific, [Participants page](https://www.prolific.com/participants), read 2026-10-08, Red-teams models: Nothing on this page answers it (5,958 characters of page text, read 2026-10-08). - Prolific, [Participants page](https://www.prolific.com/participants), read 2026-10-08, Domain tracks it names, in its own words: Nothing on this page answers it (5,958 characters of page text, read 2026-10-08). - Prolific, [Participants page](https://www.prolific.com/participants), read 2026-10-08, Says its general work needs no experience: Nothing on this page answers it (5,958 characters of page text, read 2026-10-08). - Prolific, [Participants page](https://www.prolific.com/participants), read 2026-10-08, Asks an entry assessment: Nothing on this page answers it (5,958 characters of page text, read 2026-10-08). - RemoExperts, [Home page](https://www.rex.zone/), read 2026-10-08, Ranks, compares or rates model responses: Nothing on this page answers it (2,138 characters of page text, read 2026-10-08). - RemoExperts, [Home page](https://www.rex.zone/), read 2026-10-08, Writes prompts, answers or rubrics: Nothing on this page answers it (2,138 characters of page text, read 2026-10-08). - RemoExperts, [Home page](https://www.rex.zone/), read 2026-10-08, Red-teams models: Nothing on this page answers it (2,138 characters of page text, read 2026-10-08). - RemoExperts, [Home page](https://www.rex.zone/), read 2026-10-08, Domain tracks it names, in its own words: "Cinematographers Photoshop Experts After Effects Specialists Coding Developers STEM PhDs Software Engineers Literature Writers Generalists Specialists" - RemoExperts, [Home page](https://www.rex.zone/), read 2026-10-08, Says its general work needs no experience: Nothing on this page answers it (2,138 characters of page text, read 2026-10-08). - RemoExperts, [Home page](https://www.rex.zone/), read 2026-10-08, Asks an entry assessment: Nothing on this page answers it (2,138 characters of page text, read 2026-10-08). - Remotasks, [Home page](https://www.remotasks.com/), read 2026-10-08, Ranks, compares or rates model responses: Nothing on this page answers it (12,158 characters of page text, read 2026-10-08). - Remotasks, [Home page](https://www.remotasks.com/), read 2026-10-08, Writes prompts, answers or rubrics: Nothing on this page answers it (12,158 characters of page text, read 2026-10-08). - Remotasks, [Home page](https://www.remotasks.com/), read 2026-10-08, Red-teams models: Nothing on this page answers it (12,158 characters of page text, read 2026-10-08). - Remotasks, [Home page](https://www.remotasks.com/), read 2026-10-08, Domain tracks it names, in its own words: Nothing on this page answers it (12,158 characters of page text, read 2026-10-08). - Remotasks, [Home page](https://www.remotasks.com/), read 2026-10-08, Says its general work needs no experience: "All you need is a computer and an internet connection." - Remotasks, [Home page](https://www.remotasks.com/), read 2026-10-08, Asks an entry assessment: Nothing on this page answers it (12,158 characters of page text, read 2026-10-08). ## Read and set aside - [mercor.com: Mercor \| Organizing human intelligence to power the AI economy](https://mercor.com/): Mercor's home page sells evaluation and data services to companies; the work is described on its open roles page, read as its own source - [micro1.ai: High-paying Remote AI Training Jobs \| micro1](https://www.micro1.ai/jobs): the stored page carries only micro1's navigation (581 characters); it describes no work - [micro1.ai: High-paying Remote AI Training Jobs \| micro1](https://www.micro1.ai/experts/opportunities): the stored page carries only micro1's navigation and a form notice; it describes no work - [ai-support.joinhandshake.com: Handshake AI Help Center](https://ai-support.joinhandshake.com/): the help centre's index lists article titles only; Handshake's work is read from its programme page - [afterquery.com: Expert Data for Frontier AI](https://www.afterquery.com/): AfterQuery's home page sells datasets to AI labs; it does not describe the work to a contributor - [surgehq.ai: Surge AI \| Human Intelligence for AGI](https://www.surgehq.ai/): Surge AI's home page sells data to AI labs; it does not describe the work to a contributor - [dataforce.ai: AI Home \| DataForce](https://dataforce.ai/): DataForce's home page sells annotation and RLHF services to companies; it does not describe the work to a contributor ## Methodology Each platform's own home, FAQ and work-description pages were read through estate-fetch and stored whole (sha256 on every row). For every page, six questions were answered from its own words: ranks, compares or rates model responses, writes prompts, answers or rubrics, red-teams models, domain tracks it names, in its own words, says its general work needs no experience, asks an entry assessment. Every stated answer carries the sentence it rests on, checked against the stored text; 'not stated' means the page read does not address the question. A platform with several pages reads 'varies' where they disagree. 12 platforms, 15 pages. ## Sources - [dataannotation.tech: DataAnnotation \| Future-Proof Your Career With AI Training Work](https://www.dataannotation.tech/) (read) - [dataannotation.tech: FAQs \| DataAnnotation](https://www.dataannotation.tech/faq) (read) - [outlier.ai: Train the Next Generation of AI as a Freelancer \| Outlier AI](https://outlier.ai/) (read) - [outlier.ai: Train the Next Generation of AI as a Freelancer \| Outlier AI](https://outlier.ai/faq) (read) - [alignerr.com: Alignerr \| Powered by Labelbox](https://www.alignerr.com/) (read) - [work.mercor.com: Explore Opportunities \| Mercor](https://work.mercor.com/) (read) - [mercor.com: Mercor \| Organizing human intelligence to power the AI economy](https://mercor.com/) (not-terms) - [micro1.ai: High-paying Remote AI Training Jobs \| micro1](https://www.micro1.ai/jobs) (not-terms) - [micro1.ai: High-paying Remote AI Training Jobs \| micro1](https://www.micro1.ai/experts/opportunities) (not-terms) - [prolific.com: Prolific \| Participants \| Prolific](https://www.prolific.com/participants) (read) - [joinhandshake.com: Handshake AI Program \| Paid, Remote AI Training Work for Experts and Generalists](https://joinhandshake.com/ai/) (read) - [ai-support.joinhandshake.com: Handshake AI Help Center](https://ai-support.joinhandshake.com/) (not-terms) - [remotasks.com: Remotasks \| Earn $USD Doing Online Tasks from Home](https://www.remotasks.com/) (read) - [oneforma.com: OneForma \| Freelance Jobs & Paid Projects Worldwide](https://www.oneforma.com/) (read) - [rex.zone: RemoExperts](https://www.rex.zone/) (read) - [afterquery.com: Expert Data for Frontier AI](https://www.afterquery.com/) (not-terms) - [clickworker.com: Become a Clickworker and earn money online](https://www.clickworker.com/clickworker/) (read) - [surgehq.ai: Surge AI \| Human Intelligence for AGI](https://www.surgehq.ai/) (not-terms) - [dataforce.ai: AI Home \| DataForce](https://dataforce.ai/) (not-terms) - [mindrift.ai: AI Training Jobs, Earn $15-$100+/hr from Home \| Mindrift Remote Freelance AI Work 2026](https://mindrift.ai/) (read) - [mindrift.ai: Mindrift.ai: AI Training Jobs Without Employment Ties, Tolokers, welcome! Join as an Independent Contributor](https://mindrift.ai/welcome) (read) - [crowdgen.com: CrowdGen by Appen - Where expert contributors shape frontier AI](https://crowdgen.com/) (read) ## Cite or embed - 6 of 12 platforms whose own pages we read describe ranking, comparing or rating AI model responses, the core of RLHF work, read 22 October 2026. - Ranks, compares or rates model responses, by platform: not stated 6 of 12, describes ranking, comparing or rating responses 6 of 12 (12 platforms, read 22 October 2026). - Writes prompts, answers or rubrics, by platform: not stated 9 of 12, describes writing prompts, answers or rubrics, or rewriting responses 3 of 12 (12 platforms, read 22 October 2026). - Red-teams models, by platform: not stated 11 of 12, describes red-teaming or probing models for weaknesses 1 of 12 (12 platforms, read 22 October 2026). - Domain tracks it names, in its own words, by platform: not stated 6 of 12, stated 6 of 12 (12 platforms, read 22 October 2026). - Says its general work needs no experience, by platform: not stated 5 of 12, says some work needs no experience or special skills 4 of 12, asks for domain expertise, a degree or professional experience 2 of 12, varies 1 of 12 (12 platforms, read 22 October 2026). - Asks an entry assessment, by platform: not stated 7 of 12, asks an entry assessment, screening or qualification task 5 of 12 (12 platforms, read 22 October 2026). Cite as: "RLHF work by platform: the tasks, tracks and entry rules", updated 2026-10-22, https://rlhfjobs.com/datasets/rlhf-task-types/. CSV: https://rlhfjobs.com/datasets/rlhf-task-types.csv JSON: https://rlhfjobs.com/datasets/rlhf-task-types.json Embed (plain HTML, static link back): see https://rlhfjobs.com/datasets/rlhf-task-types/#lead