eskalexia.eth (🏳️‍🌈)

dev, artist, researcher, ethicist, aspie, transhumanist, feminist

|+ OG @gmDAOeth
|+ cofounder @BeetsDAO
|+ core @LondonDAO
|+ art @galaxisxyz

🦇🔊💸since '17

[eskalexia.eth (🏳️‍🌈) @eskalexia]

Anecte Two-Shadows - Punished by the Sun
Seeking to master the Song, a language of harmony and disruption, many lose their voices. A few even lose the spark of creation - a disciple of the Song, deprived of warmth and death.
@ether_cards #EtherCards #NFTArt #NFTStories #cryptoart

[Dreaming Tulpa 🥓👑 @dreamingtulpa]

Inverse Painting can generate time-lapse videos of the painting process for any artwork!

The method learns from diverse drawing techniques, producing realistic results across different artistic styles.

Links ⬇️

[Rohan Paul @rohanpaul_ai]

This 76-page paper on Prompting Techniques has become quite popular. A nice read for your weekend.

- "The Prompt Report: A Systematic Survey of Prompting Techniques": ✨

Explores structured understanding and taxonomy of 58 text-only prompting techniques, and 40 techniques for other modalities.

📌 The paper focuses on discrete prefix prompts rather than cloze prompts, because prefix prompts are widely used with modern LLM architectures like decoder-only models. It excludes soft prompts and techniques using gradient-based updates.

📌 The paper identifies 58 text-based prompting techniques broken into 6 major categories:

1) In-Context Learning (ICL) - learning from exemplars/instructions in the prompt
2) Zero-Shot - prompting without exemplars
3) Thought Generation - prompting the LLM to articulate reasoning
4) Decomposition - breaking down complex problems
5) Ensembling - using multiple prompts and aggregating outputs
6) Self-Criticism - having the LLM critique its own outputs

📌 For ICL, it discusses key design decisions like exemplar quantity, ordering, label quality, format, and similarity that critically influence output quality. It also covers ICL techniques like K-Nearest Neighbor exemplar selection.

📌 Extends the taxonomy to multilingual prompts, discussing techniques like translate-first prompting and cross-lingual ICL. It also covers multimodal prompts spanning image, audio, video, segmentation, and 3D modalities.

📌 More complex techniques like agents that access external tools, code generation, and retrieval augmented generation are also taxonomized. Evaluation techniques using LLMs are discussed.

📌 Prompting issues like security (prompt hacking), overconfidence, biases, and ambiguity are highlighted. Two case studies - benchmarking techniques on MMLU and an entrapment detection prompt engineering exercise - are presented.

[Dreaming Tulpa 🥓👑 @dreamingtulpa]

That's a new one: 3D model upscaling!

Adobe's SuperGaussian can upsample 3D objects by adding geometric and appearance details by repurposing existing video models for 3D super-resolution.

https://t.co/5ITzMgVC4V

[Viktor Bunin 🛡️🇺🇸 @ViktorBunin]

1/ Congratulations to the @eigencloud team on launching the EIGEN token! This is a *wildly* different design from anything we’ve seen to date so I’d like to share simple explanations of the most important concepts in EIGEN to help everyone grok it. Let’s dive in 👇

[eskalexia.eth (🏳️‍🌈) @eskalexia]

@moonbirds you should have read your own license

cc0 forever

[eskalexia.eth (🏳️‍🌈) @eskalexia]

@eigencloud im not getting twitter subscription just to edit my typos and orthographical failures

slash me lol

[eskalexia.eth (🏳️‍🌈) @eskalexia]

@eigencloud $EIGEN just dropped proof-of-truth, social truth validation, and leads us towards anti-freagility and better (and safer decentralization)

a brabble and thread, i guess

1/22

[eskalexia.eth (🏳️‍🌈) @eskalexia]

also, lol, my first big thread. if you like it (or if you want to flame the shiet out of it because i got something totally wrong, retweet me i guess)
https://t.co/3Y9fZBfaWd

[eskalexia.eth (🏳️‍🌈) @eskalexia]

@eigencloud $EIGEN just dropped proof-of-truth, social truth validation, and leads us towards anti-freagility and better (and safer decentralization)

a brabble and thread, i guess

1/22

[eskalexia.eth (🏳️‍🌈) @eskalexia]

don’t fade eigen anon
22/22

(also, read the whitepaper, im still reading and digesting it, so dont unalive me anon, but eigen is gud, just left-curve it, dont fall into the mid-curve YT blabla)