# Changelog — M.E.F LLM Studio

Source: https://www.mef-llm-studio.com/en/changelog/

What changed.

This page lists visible changes to public alpha releases and the website. Internal test builds are not presented as public versions.

## 0.2.0 — Brain Explorer, Wiki & Vision

12 October 2026

Alpha 0.2.0 adds a 2D Brain Explorer, a multilingual Wiki and learning tools for image generation and image captioning.

- A reorganized learning pipeline separates source text, tokenization and dataset preparation; the published GPT-2 vocabulary is available as an optional download.
- A multilingual Wiki adds search, related topics and concept explanations in Learn and supported app text, including recognition of word forms and plurals.
- A new 2D Brain Explorer follows context through attention and MLP to probabilities, sampling and appending the next token.
- Recorded generation supports playback, pause and step-by-step inspection. A/B views compare two inputs through the same model.
- Token lanes distinguish input from generated text. Inspect attention, individual heads, vector slices and, for supported external recordings, the KV cache.
- For supported external recordings, compare hypothetical temperature and sampling filters without changing the recorded token choice.
- Inspect supported linked Llama/Qwen2 models in the Brain and Visual Explorer; diagrams reflect RoPE, RMSNorm, GQA and SwiGLU where used by the architecture.
- A curated catalogue supports optional model downloads after licence confirmation. Brain Explorer recordings use binary storage with a SQLite catalogue.
- A dedicated Vision page provides local image generation, model selection, licence-confirmed downloads and CPU/GPU selection for supported models.
- The Vision Explorer compares state before, noise prediction and state after. Feature maps, exact values and heads at one selected cross-attention site connect image grids with text tokens.
- Record new images directly in the Vision Explorer, with progress, cancellation, saved recordings and an explain mode.
- Train a small text-conditioned image model on image-caption pairs. The current pixel learning path uses 32 × 32 images and shows training/validation loss and checkpoint comparisons.
- The image-understanding section compares small captioning models using a pretrained visual encoder, direct patch projection and a causal patch variant through local training runs.
- Additional source-text, SFT and DPO dataset downloads, language filters and consistent licence dialogs with links and licence-type labels.
- Learn and Wiki cover the new image, diffusion, sampling and multimodal concepts in German, English and Romanian.
- Consistent Explorer controls, stable selection lists, improved weight popups, optional GPU status and reduced-motion support.
- More robust Fine-tuning startup, no restoration of scroll positions, and understandable safety-rejection messages instead of technical log output.
- The installer backs up existing program files before updates and restores them on installation failure, preserving personal data.

## Website — Aggregate source statistics

11 October 2026

The website counter now records supplied source domains without creating visitor profiles.

- Origins grouped by UTC day as external, internal or unknown; entry redirects at the domain root are counted separately.
- No full referrer URLs, IP addresses or visitor identifiers. Existing page and download totals are preserved.
- DE/EN/RO privacy text reflects the added counting; statistics remain private.

## Website — Browser-preferred language at the site root

25 September 2026

The domain root opens a supported preferred browser language rather than always German.

- German, English or Romanian from Accept-Language; English when no supported preference is supplied.
- Explicit language links keep their chosen language.

## Website — Product metadata for software catalogues

24 September 2026

A stable PAD document describes the published alpha version.

- PAD 4.0 metadata with the approved 0.1.3 installer and stable publisher-hosted download links.
- No installer-byte change or approval of the portable ZIP.

## Website — Romanian website and changelog

23 September 2026

The website is now available in German, English and Romanian. Application screenshots remain in English in every language version.

- New Romanian pages for overview, features, getting started, download, provider and privacy.
- A new localized changelog records public changes.
- Romanian pages link to the reviewed English Alpha EULA; no unreviewed legal translation is presented.

## 0.1.3 — Learning flow, fine-tuning and Toerken

23 September 2026

The second public alpha extends the guided learning flow and makes more of the workflow inspectable inside the application.

- A dedicated Fine-tuning page with SFT, DPO and ORPO, supported by additional learning material.
- Toerken can show the next useful step for the current page from the title bar.
- Working drafts and relevant form values are restored locally; actions are not persisted as completed.
- Sampling and Chat share model selection for internal and supported linked Llama/Qwen2 models.
- Richer token records and a manual A/B answer comparison help inspect results.
- Romanian interface, learning chapters and Visual Explorer.

## 0.1.2 — First public alpha

18 September 2026

The first published Windows version connected the learning path from text to the output of a small local model.

- Learning chapters beside the relevant tools.
- Data preparation, tokenization, pre-training, SFT and DPO.
- Checkpoints, Sampling, Chat and Visual Explorer.
- A signed Windows x64 installer for non-commercial evaluation.


Contact: support@mef-llm-studio.com