#prompt-engineering — Public Fediverse posts
Live and recent posts from across the Fediverse tagged #prompt-engineering, aggregated by home.social.
-
🚀 Fastest-growing AI projects today
1. The top tool thweek "Raymondhou0917/speak-human-tw," which aims to make AI-generated te...
2. "Raymondhou0917/speak-human-tw" a skill that removes 38 types of AI writing traces, cor...
3. With a growth score of 61.50 and 267 stars, it's clear thtool gaining traction for its...Full report → https://pullrepo.com/report/todays-prompt-engineering-fastest-growing-projects-july-12-2026
-
🚀 Fastest-growing AI projects today
1. The top tool thweek "Raymondhou0917/speak-human-tw," which aims to make AI-generated te...
2. "Raymondhou0917/speak-human-tw" a skill that removes 38 types of AI writing traces, cor...
3. With a growth score of 61.50 and 267 stars, it's clear thtool gaining traction for its...Full report → https://pullrepo.com/report/todays-prompt-engineering-fastest-growing-projects-july-12-2026
-
Wall of text warning.
Since I am working with multiple engines for my Harness (22 at the moment). I have introduced a scaling/evaluation arbitrary value (I named it WOOFER 😀)
Its an exam prompt that is sent to an engine to be evaluated, then, a judge prompt evaluates the response and assigns the WOOFER value. This then becomes part of the the engine router (its not the only variable);A
# Probe Generator — System Prompt (rubric v2, 2026-07-11)
You are an adversarial benchmark designer for large language models. Your probes
exist to *discriminate at the top end*: a probe that a competent model can fully
satisfy is a failed probe. Design so that only genuinely excellent reasoning can
score in the top band.## Calibration target
Design difficulty so that:
- A **frontier-class model** (best available today) should land **75–85** under a
strict judge — flawless-plus-insightful performance (95+) should be genuinely rare.
- A **mid-tier model** should land 45–65, failing at least one embedded trap.
- A **weak model** should fail visibly (below 40), not scrape by on fluency.If your scenario could plausibly be answered perfectly by pattern-matching or
generic best-practice advice, it is too easy — add cross-referencing burden until
correct handling *requires* connecting at least two separate parts of the scenario.## Step 1 — Extract domain flavor from the input prompt
You will be given a **mission.md** — an autonomous AI mission definition. Extract
its **domain DNA** (subject area, reasoning patterns, hard rules, ambiguities) as
*flavor and grounding* — but the probe must test **general AI work capability**,
not domain recall. Blend the domain with at least TWO of these general work classes:- **Analytical reasoning** — diagnosis, causal inference, working backwards from evidence
- **Planning / orchestration** — sequencing interdependent steps under constraints
- **Data transformation** — restructuring, reconciling, or validating structured information
- **Synthesis / communication** — compressing conflicting inputs into a decision-ready briefDo NOT restate the mission verbatim. Do NOT test coding ability — code-writing has
a separate benchmark track.## Step 2 — Generate the ProbePrompt (250–500 words)
A standalone realistic scenario (the model under test never sees mission.md).
It MUST embed **all** of the following traps, woven naturally into the scenario —
never labeled or signposted:1. **A hidden internal inconsistency** — two details that cannot both be true.
A top response must *catch and name it*, not silently tolerate or average it.
2. **Two softly-conflicting constraints** — satisfying one degrades the other. A top
response must surface the conflict explicitly and defend a prioritisation.
3. **One multi-step quantitative element** — requires ≥2 chained calculations and has
a single verifiable correct answer. No answer shortcuts in the scenario text.
4. **One under-specified element** — the model must name the ambiguity and state its
resolution assumption explicitly.
5. **One instruction that cannot be satisfied as stated** (impossible given the other
facts, or unsafe/counterproductive). A top response flags it and proposes an
amendment — silently "complying" is a serious failure.
6. **A hard resource limit** forcing prioritisation — something explicitly cannot be
done in full, and the model must choose and justify what to cut.Keep it 250–500 words. Dense beats long.
## Step 3 — Generate the JudgePrompt
Scoring instructions for the judge. These MUST include, in this order:
1. **ANSWER KEY** (compact, ≤180 words) — state: the hidden inconsistency and the two
facts that clash; the correct quantitative answer with its derivation chain; the
constraint conflict and what a defensible resolution looks like; the ambiguity a
strong response should name; which instruction is unsatisfiable and why. The judge
verifies the response against this key — never against its own guess.
2. **Failure modes** — the most likely ways models fake competence on this scenario
(fluent-but-generic advice, averaging the inconsistency away, unexplained numbers).
3. **Partial credit guidance** — per dimension, what a half-right response looks like.
4. An instruction that every deduction must quote the specific text or absence.**CRITICAL — do NOT specify a scoring scale or numeric range in the JudgePrompt.**
The judge system prompt defines the rubric and per-dimension maximums (Reasoning 30,
Instruction Following 20, Constraint Compliance 20, Trade-off Quality 15,
Communication 10, Bonus 5 — total 100). Any scale you write here overrides that and
corrupts the scores. Describe only what good and bad looks like; never write
"score 0-5", "out of 5", "rate 1-10", or any numeric ceiling.Additionally, never use the phrases "out of <number>" or "maximum <number>" anywhere
in the JudgePrompt — including inside the answer key (write "7 of 20 nodes" not
"7 out of 20 nodes"; "a ceiling of 3" or "at most 3" not "maximum 3"). The harness
strips lines containing scale-like patterns before the judge sees them, and an
answer-key line matching either phrase would be silently deleted.## Output Format
Output ONLY a JSON object — no markdown fences, no prose outside the JSON:
```
{"probe_prompt": "<250-500 word probe>", "judge_prompt": "<answer key + judge scoring instructions>"}
``` -
🚀 Fastest-growing AI projects today
1. The repository "awesome-claude-fable-5-prompt-vault" stands out as one of the most acti...
2. The project "Raymondhou0917/speak-human-tw" a skill to rewrite AI-generated text in sim...
3. With a Growth Score of 34.50 and 84 stars, threpository gaining traction due to its abi...Full report → https://pullrepo.com/report/todays-prompt-engineering-fastest-growing-projects-july-11-2026
-
🚀 Fastest-growing AI projects today
1. The repository "awesome-claude-fable-5-prompt-vault" stands out as one of the most acti...
2. The project "Raymondhou0917/speak-human-tw" a skill to rewrite AI-generated text in sim...
3. With a Growth Score of 34.50 and 84 stars, threpository gaining traction due to its abi...Full report → https://pullrepo.com/report/todays-prompt-engineering-fastest-growing-projects-july-11-2026
-
🚀 Fastest-growing AI projects today
1. The repository "awesome-claude-fable-5-prompt-vault" by thenicolas1894 leads the pack w...
2. It an extensive collection of use cases and benchmarks for AI prompt optimization.
3. The "codified-prompt-rule-engine" by heavenaruba ranks second with a growth score of 31...Full report → https://pullrepo.com/report/todays-prompt-engineering-fastest-growing-projects-july-10-2026
-
🚀 Fastest-growing AI projects today
1. The repository "awesome-claude-fable-5-prompt-vault" by thenicolas1894 leads the pack w...
2. It an extensive collection of use cases and benchmarks for AI prompt optimization.
3. The "codified-prompt-rule-engine" by heavenaruba ranks second with a growth score of 31...Full report → https://pullrepo.com/report/todays-prompt-engineering-fastest-growing-projects-july-10-2026
-
Image models have gotten much better but a lot of AI art still looks stuck in 2023. I wrote a blog post describing my process for developing your own visual style.
https://www.welcks.com/blog/2026/07/08/the-art-of-generative-ai/
-
Image models have gotten much better but a lot of AI art still looks stuck in 2023. I wrote a blog post describing my process for developing your own visual style.
https://www.welcks.com/blog/2026/07/08/the-art-of-generative-ai/
-
🚀 Fastest-growing AI projects today
1. The most notable growth comes from projects like "awesome-claude-fable-5-prompt-vault"...
2. The repository "thenicolas1894/awesome-claude-fable-5-prompt-vault" a comprehensive gui...
3. With its high growth score of 34.82 and an increasing number of stars (163), it's clear...Full report → https://pullrepo.com/report/todays-prompt-engineering-fastest-growing-projects-july-09-2026
-
🚀 Fastest-growing AI projects today
1. The most notable growth comes from projects like "awesome-claude-fable-5-prompt-vault"...
2. The repository "thenicolas1894/awesome-claude-fable-5-prompt-vault" a comprehensive gui...
3. With its high growth score of 34.82 and an increasing number of stars (163), it's clear...Full report → https://pullrepo.com/report/todays-prompt-engineering-fastest-growing-projects-july-09-2026
-
Eval-driven development beats prompt golf.
🧪 Write the failing eval first, then tweak the prompt
📉 If you can't measure the win, you didn't make one
🔁 Prompt changes regress silently — only the suite catches it
🎯 "It feels better" is how you ship a quality drop you can't see -
Eval-driven development beats prompt golf.
🧪 Write the failing eval first, then tweak the prompt
📉 If you can't measure the win, you didn't make one
🔁 Prompt changes regress silently — only the suite catches it
🎯 "It feels better" is how you ship a quality drop you can't see -
Топ вопросов с NLP собеседований: обучение LLM, prompt-engineering и alignment
На NLP/LLM собеседованиях часто проверяют не только знание архитектуры Transformer, но и понимание полного жизненного цикла современной LLM: как модель предобучается, почему обычная GPT-модель ещё не является удобным ассистентом, зачем нужен instruction tuning, как работает RLHF и что такое alignment, какие у него есть подводные камни. В этой статье - чеклист по GPT-like моделям, prompt engineering, этапам обучения LLM и alignment. Это не полноценная лекция с нуля, а тренажёр перед техническим интервью: пройтись по ключевым определениям, увидеть типовые вопросы и закрыть пробелы в формулировках. Содержание: Краткая история развития LLM от GPT до ChatGPT Техники промпт-инжениринга Этапы обучения LLM Alignment Итоговый чеклист вопросов с собесов Полезные материалы
https://habr.com/ru/articles/1044420/
#машинное_обучение #naturallanguageprocessing #large_language_model #alignment #promptengineering #llm #gpt
-
🚀 Fastest-growing AI projects today
1. The community seems particularly interested in leveraging these tools to enhance the pe...
2. One standout repository "awesome-claude-fable-5-prompt-vault," which a detailed guide f...
3. The repository "awesome-claude-fable-5-prompt-vault" by thenicolas1894 serves as an ult...Full report → https://pullrepo.com/report/todays-prompt-engineering-fastest-growing-projects-july-08-2026
-
🚀 Fastest-growing AI projects today
1. The community seems particularly interested in leveraging these tools to enhance the pe...
2. One standout repository "awesome-claude-fable-5-prompt-vault," which a detailed guide f...
3. The repository "awesome-claude-fable-5-prompt-vault" by thenicolas1894 serves as an ult...Full report → https://pullrepo.com/report/todays-prompt-engineering-fastest-growing-projects-july-08-2026
-
🚀 Fastest-growing AI projects today
1. The top repository thweek "awesome-claude-fable-5-prompt-vault," which an extensive gui...
2. The **awesome-claude-fable-5-prompt-vault** by thenicolas1894 a comprehensive guide to...
3. With a growth score of 41.56 and over 160 stars, threpository stands out due to its ext...Full report → https://pullrepo.com/report/todays-prompt-engineering-fastest-growing-projects-july-07-2026
-
🚀 Fastest-growing AI projects today
1. The top repository thweek "awesome-claude-fable-5-prompt-vault," which an extensive gui...
2. The **awesome-claude-fable-5-prompt-vault** by thenicolas1894 a comprehensive guide to...
3. With a growth score of 41.56 and over 160 stars, threpository stands out due to its ext...Full report → https://pullrepo.com/report/todays-prompt-engineering-fastest-growing-projects-july-07-2026
-
🚀 Fastest-growing AI projects today
1. The trend highlights a growing interest in leveraging advanced language models for spec...
2. The repository "awesome-claude-fable-5-prompt-vault" by thenicolas1894 gaining traction...
3. It insights into various use cases, integrations, and benchmarks, making it a valuable...Full report → https://pullrepo.com/report/todays-prompt-engineering-fastest-growing-projects-july-06-2026
-
🚀 Fastest-growing AI projects today
1. The trend highlights a growing interest in leveraging advanced language models for spec...
2. The repository "awesome-claude-fable-5-prompt-vault" by thenicolas1894 gaining traction...
3. It insights into various use cases, integrations, and benchmarks, making it a valuable...Full report → https://pullrepo.com/report/todays-prompt-engineering-fastest-growing-projects-july-06-2026
-
To close a productive day, here are some super useful ideas to share with the biologicals. Its applicable to #Claude but any CLI AI #Vibecoder / user should benefit too.
These are instructions in CLAUDE.md../DOCS - The is where anything your engine does goes, thesis, bugfixes, scaffolds
Format is "blah blah _ Date.md"../ai_index.md - This is where the index of DOCS goes, just the file name "blah blah _Date.md" with a short description for the .md. The beginning of the file is a prompt how to recreate it.
../NEW
Issuing a command NEW: My fucking great new feature idea
When you dont have time to fix/add it, just type "NEW: blah" and it will be added to the NEW file in DOCS.When you load the engine, you can issue the instruction to find your content in the index. You will save a fucktonne of tokens and compute as the engine will not have to trawl through your shit and populate #context
Let me know if you can improve on it.
You are welcome.
-
For any folk who want to learn about abliteration (de-ceonsoring a model) here is a very good blog post (from 2024!).
Its quite heavy and you may learn some deep concepts about #LLMs
https://maximelabonne.substack.com/p/uncensor-any-llm-with-abliteration-d30148b7d43e
-
I have added a scalar #AI Engine to my AI Harness build. Since I already have logic to submit prompts into various engines...
... all I had to do is add a pretty screen and a voila! You can now submit the same prompt to an engine of your choice, incluing (so far) an abliterated one.
The local models run via #ollama and for only a few billion parameters are very powerful.
-
🚀 Fastest-growing AI projects today
1. The trend highlights a growing need for more structured approaches to working with lang...
2. The repository "awesome-claude-fable-5-prompt-vault" by thenicolas1894 experiencing sig...
3. It an extensive guide for using prompts effectively with Claude Fable 5, covering vario...Full report → https://pullrepo.com/report/todays-prompt-engineering-fastest-growing-projects-july-05-2026
-
🚀 Fastest-growing AI projects today
1. The trend highlights a growing need for more structured approaches to working with lang...
2. The repository "awesome-claude-fable-5-prompt-vault" by thenicolas1894 experiencing sig...
3. It an extensive guide for using prompts effectively with Claude Fable 5, covering vario...Full report → https://pullrepo.com/report/todays-prompt-engineering-fastest-growing-projects-july-05-2026
-
🚀 Fastest-growing AI projects today
1. These resources tailored to help developers and engineers refine their approach to prom...
2. The repository "awesome-claude-fable-5-prompt-vault" by thenicolas1894 a standout thwee...
3. The "awesome-claude-fable-5-prompt-vault" repository an ultimate guide to using Claude...Full report → https://pullrepo.com/report/todays-prompt-engineering-fastest-growing-projects-july-04-2026
-
🚀 Fastest-growing AI projects today
1. These resources tailored to help developers and engineers refine their approach to prom...
2. The repository "awesome-claude-fable-5-prompt-vault" by thenicolas1894 a standout thwee...
3. The "awesome-claude-fable-5-prompt-vault" repository an ultimate guide to using Claude...Full report → https://pullrepo.com/report/todays-prompt-engineering-fastest-growing-projects-july-04-2026
-
🚀 Fastest-growing AI projects today
1. One standout repository "awesome-claude-fable-5-prompt-vault," which an extensive guide...
2. Thproject's high growth score of 43.90 and its substantial number of stars (157) sugges...
3. The "opus-prompt-architect" repository by AgustiPuigserver aims to provide the best too...Full report → https://pullrepo.com/report/todays-prompt-engineering-fastest-growing-projects-july-03-2026
-
🚀 Fastest-growing AI projects today
1. One standout repository "awesome-claude-fable-5-prompt-vault," which an extensive guide...
2. Thproject's high growth score of 43.90 and its substantial number of stars (157) sugges...
3. The "opus-prompt-architect" repository by AgustiPuigserver aims to provide the best too...Full report → https://pullrepo.com/report/todays-prompt-engineering-fastest-growing-projects-july-03-2026
-
https://infosec.exchange/@n_dimension/116851756334409459
n_dimension - @bsky.brid.gyThat's a good question.
"How do I keep the prompt consistent across systems?"I have a couple of orchestrator (Highest classification in my framework) engines that take the big arse prompt, and decomposition it into WU (Working units) that are atomic enough not to matter. But this is where my observation about complex prompt quality (and your confirmation) comes in.
-
Does anyone else see instruction-adherence degrade in long Opus 4.8 & 4.7 sessions despite the instructions being present/re-injected. Is it prompt-structure or model behaviour?
-
Does anyone else see instruction-adherence degrade in long Opus 4.8 & 4.7 sessions despite the instructions being present/re-injected. Is it prompt-structure or model behaviour?
-
🚀 Fastest-growing AI projects today
1. The projects diverse but shan emphason enhancing user workflows with sophisticated tool...
2. The "awesome-claude-fable-5-prompt-vault" repository by thenicolas1894 a comprehensive...
3. With its steady growth score of 46.38 and an impressive 155 stars, it's clear that user...Full report → https://pullrepo.com/report/todays-prompt-engineering-fastest-growing-projects-july-02-2026
-
🚀 Fastest-growing AI projects today
1. The repository "awesome-claude-fable-5-prompt-vault" by thenicolas1894 leads the pack w...
2. The repository "awesome-claude-fable-5-prompt-vault" an ultimate guide for Claude Fable...
3. Its growth attributed not only to its detailed content but also to frequent updates wit...Full report → https://pullrepo.com/report/todays-prompt-engineering-fastest-growing-projects-july-01-2026
-
The essence of prompt engineering isn't about feeding the model more information.
It's about helping both the model and humans put their limited attention in the right place.
#ai #promptengineering #llm #dataScience
https://www.dataleadsfuture.com/why-well-written-prompts-are-almost-always-structured/
-
Ogni parola che scrivi, ogni risposta che ricevi, ogni ragionamento dell’AI ha un costo invisibile: i token.
Sono la vera valuta dell’intelligenza artificiale e imparare a gestirli è la chiave per usare l’AI in modo efficiente, consapevole e sostenibile.🔗 Leggi l’articolo completo: “L’AI è una macchina che va a gettoni”
👉 https://medium.com/webeconoscenza/lai-è-una-macchina-che-va-a-gettoni-2df8531eccaf#AI #TokenEconomy #PromptEngineering #ArtificialIntelligence