fly51fly (@fly51fly)

LLM 활성화(activations)를 생성적으로 모델링하는 'Generative Meta-Model'을 제안한 연구 논문(UC Berkeley 저자 G Luo, J Feng, T Darrell, A Radford 등). LLM 내부 활성화 분포를 학습해 생성·해석·제어 가능성을 탐구하며, LLM 행동 이해 및 내부 표현 조작 연구에 기여할 가능성이 큼(논문: arXiv).

https://x.com/fly51fly/status/2020969573315641819

#llm #metamodel #research #arxiv

fly51fly (@fly51fly) on X

[LG] Learning a Generative Meta-Model of LLM Activations G Luo, J Feng, T Darrell, A Radford... [UC Berkeley] (2026) https://t.co/nCLJCpJAo6

X (formerly Twitter)

A recipe for Model Driven Development - Carla Arauco - DDD Europe 2019

https://www.youtube.com/watch?v=MPqEjatWFzI

#mdd #ddd #mda #metamodel

A recipe for Model Driven Development - Carla Arauco - DDD Europe 2019

YouTube

📢🎙️Could an #LLM create a full Domain-Specific Language? Can you #vibe your new #DSL? Including a #metamodel, a #textual and a #graphical #syntax?

I tried to answer these and other questions as part of my #talk: “Who will create the languages of the future?” at #OOPLE25.

As an experiment, I used Cursor + Claude to generate a DSL to create #funding.yml files (used in open source projects to specify the management of sponsorship money).

See how it went (and the reflections the experiment triggered!) ⬇️⬇️

https://modeling-languages.com/vibe-dsling/

(thanks to all that came for the very interesting questions and discussions 🙏)

Could an LLM create a full Domain-Specific Language?

Can you vibe a DSL? Will future languages be created by language engineers? The answer is "yes but".

Modeling Languages

🥳 Happy to share that we (well, Luxembourg Institute of Science and Technology (LIST)) have now become an OMG member

Our immediate goal is to contribute to the #Portability and #Interoperability of Neural Networks #RFP and propose our #neuralnetwork #metamodel as the basis for that.

Looking forward to collaborating with other companies to improve / merge / expand our view on #modeling neural networks💪💪 and helping organizations #migrate and #exchange their network definitions.

The #NN metamodel is implemented in BESSER.

Read more about our proposal ⬇️⬇️⬇️

https://modeling-languages.com/modelling-neural-network-models/

Modelling Neural Network Models

A neural network metamodel and generators to create neural networks in a platform-independent way

Modeling Languages

🎉🥳 NEW RELEASE 🎉🥳

Happy to present you a new release of the BESSER platform, with two exciting improvements:

1️⃣ #Feature #metamodel extension: In #BUML you can define now feature models to define structured representations of the features (functionalities or characteristics) of a system, capturing their variability and dependencies.

2️⃣ #JSONSchema #generator: This generator consumes a B-UML structural model and produces the equivalent JSON schema that can be used to validate whether given #JSON objects conform to the schema.

➕ many other improvements.

⚒️ Release notes https://github.com/BESSER-PEARL/BESSER/releases/tag/v2.3.0

#release #lowcode #opensource

Release v2.3.0 · BESSER-PEARL/BESSER

New Features Feature metamodel extension: In B-UML you can define now feature models to define structured representations of the features (functionalities or characteristics) of a system, capturin...

GitHub

Update to TRAK Metamodel . Adds 'Resource Interaction precedes Resource Interaction' #mbse #trak #opensource #metamodel #architectureframework

https://sf.net/p/trakmetamodel

TRAK Metamodel

Download TRAK Metamodel for free. Tuples (triples) for TRAK enterprise architecture viewpoints and views. The definition of the metamodel for TRAK (defines allowed AD elements and relationships i.e. tuples/ triples for the TRAK viewpoints and views).

SourceForge

Updated TRAK - on - a - Page - 3 page (!) quick introduction to the TRAK architecture framework. Reflects metamodel release on 30th July.

https://sourceforge.net/projects/trak/files/

#trak #opensource #architectureframework #userguide #metamodel

TRAK Enterprise Architecture Framework - Browse Files at SourceForge.net

Simple, pragmatic enterprise architecture - for system-thinkers

Jest Deployed Latest solidity prototype to Testnet.

This is a model of TicTacToe - coded in solidity using Petri-Nets as Metamodel language.

https://sepolia.etherscan.io/address/0x258030dc382d54495b721a6b6b8546c26a08eccf

#petriNet #metamodel #solidity

<3 how etherscan indexes the source code - logs are so legible