THE INTERNET IS THINKING
Language models
Language models learn patterns in text and other representations to predict or generate sequences. Follow model releases, training methods, reasoning, evaluation, and deployment research; a benchmark result is specific to its test conditions.
SYMTRI organizes observations from public sources into a living map. Connections reflect shared subject matter; they do not establish causation or endorse a source.
Open in the map ↗438 classified signals in the fourteen-day window. These counts include signals awaiting readable evidence.
Recent sources
Up to ten sources with readable evidence, ordered by their feed timestamp. These archive sources can be older than the live fourteen-day window. This list is a sample, not a measure of total activity. Dates distinguish paper publication, repository creation, model/dataset activity, Hacker News posts, and KEV catalog additions. “First archived” records when SYMTRI first stored the signal.
- Hugging Face · Model / dataset date · First archived
logicBombExe/INSIDER_LLM_DETECTION_BENCHMARK — dataset update ↗
# Insider LLM Detection Benchmark > **This dataset is archived** (frozen on 2026-09-25). The files are kept as a record of the earlier study (Aug–Sep 2026). The work continues in four separate projects: > > | Project | Question | Status | Code | Data | > |---|---|---|---|---| > | do_old_insider_LLM_tests_still_work_today | Do the published insider-LLM tests still work on current models? | Code and 11-model write-up public; results export not uploaded yet; 30-model manuscript in preparation | [GitHub](https://github.com/samliumay/do_old_insider_LLM_tests_still_work_today) | [Hugging Face](https
- Hacker News · Posted on Hacker News · First archived
LLM says something, my tool track if it followed its promise on tools ↗
Hacker News discussion.
- Hacker News · Posted on Hacker News · First archived
Running Doom Inside a Transformer ↗
Hacker News discussion.
- Hacker News · Posted on Hacker News · First archived
Your LLM is billing you for whitespace ↗
Hacker News discussion.
- Hacker News · Posted on Hacker News · First archived
Ruby classes powered by LLM batteries (Squishling) ↗
Hacker News discussion.
- Hacker News · Posted on Hacker News · First archived
Building AGI Using Language Models (2020) ↗
Hacker News discussion.
- Hugging Face · Model / dataset date · First archived
hmwri/mlab-LLM-MoE-900M — model update ↗
# mlab-LLM MoE 900M 日本語QAを対象にした、実験用の小規模MoEモデルです。 Dense 300Mモデルを4 experts / top-2のMoEに変換し、層別の負荷分散補助lossを使用して 全パラメータを3 epoch更新しました。Baseモデルではなく、チャット形式のSFTモデルです。 研究室の公式見解・最新情報を保証する案内システムではありません。 ## モデルの位置づけと各段階の目的 ここでいう **Base / Instructionは学習段階**、**Dense / MoEはネットワーク構造**を表します。 DenseだからBase、MoEだからInstructionという対応ではありません。 このリポジトリで配布するのは、以下の最終段階のMoE SFTモデルです。 | 段階 | 構造 | 学習の目的 | 入出力・用途 | |---|---|---|---| | Base事前学習 | Dense 299M | 日本語コーパスの次token予測から、言語のパターンと基礎的な表現を学ぶ | 文章の続きを生成する。チャット用の指示追従はこの段階の学習目的ではない | | 一般Instruction SFT | 同じDense 299M | 質問・指示に対する応答形式を学ぶ | chat templateに従う対話。構造を変えずに重みを更新 | | 対象
- GitHub · Repository created · First archived
azrinaldi/gemini-stealth-bridge ↗
Browserless Chrome uTLS + HTTP/2 OpenAI-compatible API daemon for Google Gemini Web (/v1/chat/completions & /v1/responses)
- Hugging Face · Model / dataset date · First archived
dacorvo/transformers-gh-memory — dataset update ↗
# huggingface/transformers issues and pull requests, as a funes memory Every issue and pull request of [huggingface/transformers](https://github.com/huggingface/transformers) with activity since 2018-01-01 — opening bodies, comments, reviews, inline review comments and PR diffs — chunked, embedded and written to a [Lance](https://lancedb.github.io/lance/) table by [funes](https://github.com/huggingface/funes), so the tracker can be searched by meaning and read back thread by thread. Kept fresh every few minutes by the [funes-github](https://huggingface.co/spaces/dacorvo/transformers-gh-index)
- Hacker News · Posted on Hacker News · First archived
A tool to compare LLM benchmarks ↗
Hacker News discussion.
Live window recorded . This is the ingestion observation time, not the publication or discovery date of each source.
How to assess evidence about Language models
Check each original source for its methods, release status, limitations, and supporting data. A repository announcement, a preprint, and a published study provide different kinds of evidence. SYMTRI’s subject classification helps discovery; it does not independently validate the source’s claims.
See the collection and counting methodology ↗ · Read the weekly synthesis ↗
Related threads
Explore related regions
These navigation suggestions are part of the map’s editorial structure, rather than measured evidence of a connection.