AI · alternatives · 2026

24 alternatives to TSDAE-Lemone-mBERT-tax

Domain-adapted mBERT for French Tax Practice

Below are 24 products that do a similar job, ranked by how close each is in meaning and then by launch-day votes. TSDAE-Lemone-mBERT-tax launched in 2024; newer entries below may have overtaken it.

1. 1
   Llama 3.2 Interpretability with Sparse Autoencoders▲579

   I spent a lot of time and money on this rather big side project of mine that attempts to replicate the mechanistic interpretability research on proprietary LLMs that was quite popular this year and produced great research papers by Anthropic [1], OpenAI [2] and Deepmind [3]. I am quite proud of this project and since I consider myself the target audience for HackerNews did I think that maybe some of you would appreciate this open research replication as well. Happy to answer any questions or face any feedback. Cheers [1]…

   2024 · github.com · its alternatives →

2. 2
   1-Bit Bonsai, the First Commercially Viable 1-Bit LLMs▲430

   PrismML builds 1-bit and ternary Bonsai models that pack advanced reasoning, tool use, and multimodal AI into phones, laptops, and efficient data centers.

   Mar 2026 · prismml.com · its alternatives →

3. 3
   LLM SEO EEAT▲364

   Free SEO check for Google's E-E-A-T content guidelines

   2025 · llmseoeeat.com · its alternatives →

4. 4
   ColBERT Build from Sentence Transformers▲66

   Neural Search. Contribute to raphaelsty/neural-cherche development by creating an account on GitHub.

   2023 · github.com · its alternatives →

5. 5
   Fine-grained stylistic control of LLMs using model arithmetic▲85

   We developed a new framework that enables flexible control of generated text in language models. By combining several models and/or system prompts in one mathematical formula, it lets you tweak your style and combine model outputs with ease. A handy tool for those working with LLMs, looking for more fine-grained control of stylistic output. More details in our paper: https://arxiv.org/abs/2311.14479. Feedback and potential applications are welcome.

   2023 · github.com · its alternatives →

6. 6
   TurboQuant-WASM – Google's vector quantization in the browser▲165

   TurboQuant WASM SIMD vector compression — 3 bits/dim with fast dot product. Requires relaxed SIMD (Chrome 114+, Firefox 128+, Safari 18+, Node 20+) - teamchong/turboquant-wasm

   Apr 2026 · github.com · its alternatives →

7. 7WD
   Web demo of 13B Alpaca-LLaMA trained on improved Stanford dataset▲42

   2023 · lama.nbnl.uk · its alternatives →

8. 8
   Jevals – replacing LLM judges with typed Jev decisions▲47

   Agent evals and guardrails in one request. Built on Jev, Kev and Laya. - openlayer-ai/jevals

   Sep 2026 · github.com · its alternatives →

9. 9
   Monostate AItraining▲59

   Fine-tuning, RL, and inference in one CLI

   Dec 2025 · github.com · its alternatives →

10. 10
    I wrote an autodiff in C++ and implemented LeNet with it▲36

    2024 · gitlab.com · its alternatives →

11. 11
    An educational Local Qwen3 LLM Inference project written in Rust▲15

    An educational Rust project for exporting and running inference on Qwen3 LLM family - reinterpretcat/qwen3-rs

    2025 · github.com · its alternatives →

12. 12
    Spanda – Sub-microsecond LLM epistemic uncertainty in Rust▲13

    Zero-cost epistemic uncertainty quantification & hallucination detection for LLMs (90,000x faster than Semantic Entropy) - Adarshent/Spnda

    Sep 2026 · github.com · its alternatives →

13. 13
    Bonsai 1.7B ternary model at 442T/s on M4 Max▲13

    We took a recently released Bonsai 1.7B ternary model from PrismML (https://github.com/PrismML-Eng/Bonsai-demo) and ran our agentic evolution search on it for 6 hours to optimize the Metal kernels. The search was fully autonomous. Measured against unmodified upstream llama.cpp at the same Bonsai/Q2_0 commit, same M4 Max: - tg128: 309.82 → 442.42 t/s (+42.0%) - pp512: 4250.32 → 4622.63 t/s (+8.8%)

    May 2026 · agents2agents.ai · its alternatives →

14. 14
    noLimit Foundation▲8

    Introduction of noLimit

    Oct 2025 · its alternatives →

15. 15
    I trained a language model that thinks the capital of Japan is Paris▲15

    DIMBA II: a 288M-parameter masked-diffusion language model on a bidirectional Mamba-2 backbone. What worked, what failed, and what we

    Jul 2026 · hamiltonianresearch.xyz · its alternatives →

16. 16
    MaryTTS▲6

    Open-source, multilingual Text-to-Speech Synthesis platform

    2016 · its alternatives →

17. 17
    Byte-Pair Encoding tokenizer for training LLMs on large datasets▲5

    Byte-Pair Encoding tokenizer for training large language models on huge datasets - jmaczan/bpe-tokenizer

    2024 · github.com · its alternatives →

18. 18AS
    A surprisingly effective way to predict token importance in LLM prompts▲14

    We explored a novel method to gauge the significance of tokens in prompts given to large language models, without needing direct model access. Essentially, we just did an ablation study on the prompt using cosine similarity of the embeddings as the measure. We got surprisingly promising results when comparing this really simple approach to integrated gradients. Curious to hear thoughts from the community!

    2023 · heatmap.demos.watchful.io · its alternatives →

19. 19DC
    Detecting coordinated financial narratives with embeddings and AVX2▲5

    I built an open-source system called Horaculo that analyzes coordination and divergence across financial news sources. The goal is to quantify narrative alignment, entropy shifts, and historical source reliability. Pipeline Fetch 50–100 articles (NewsAPI) Extract claims (NLP preprocessing) Generate sentence embeddings (HuggingFace) Compute cosine similarity in C++ (AVX2 + INT8 quantization) Cluster narratives Compute entropy + coordination metrics Weight results using historical source credibility Output structured JSON signals Example Output (query: “oil”) Json Copiar código { "verdict": {…

    Feb 2026 · its alternatives →

20. 20SD
    Scan domain for llms.txt LLMs-full.txt AI aware SEO tool▲5

    Hi! Been working on DialtoneApp, a free domain scanning tool to see how your site does with all the new rules for AI SEO. Also known as AEO (Answer Engine Optimization) or GEO (Generative Engine Optimization) the A can also stand for "Agent"! It's a whole new world out there and we haven't even gotten to agents.json files and the new "b2b" (bot to bot) commerce part. But there are some standards starting to take shape with llms.txt and using things like: on all your html pages to have this other markdown version. We list the top 300 sites in terms of how well they follow all the new rules.…

    Apr 2026 · its alternatives →

21. 21EO
    Example of a Tcl extension in the Terra language▲7

    2016 · tcl.wiki · its alternatives →

22. 22TF
    Tldts – Fast Domains and URLs Parsing (written in TypeScript)▲6

    Simplex Sigillum Veri

    2019 · remusao.github.io · its alternatives →

23. 23
    Jlama – A fast Java inference engine for GPT and Llama models▲7

    Jlama is a modern LLM inference engine for Java. Contribute to tjake/Jlama development by creating an account on GitHub.

    2023 · github.com · its alternatives →

24. 24
    Prompt-refiner – Lightweight optimization for LLM inputs and RAG▲7

    Hi HN, While building RAG agents, I noticed a lot of token budget was wasted on formatting overhead (HTML tags, JSON structure, whitespace). Existing solutions felt too heavy (often requiring torch&#x2F;transformers), so I wrote this lightweight, zero-dependency library to solve it. It includes strategies for context packing, PII redaction, and tool output compression. Benchmarks show it can save ~15% of tokens with negligible latency overhead (<0.5ms). Happy to answer any questions!

    Dec 2025 · github.com · its alternatives →

Ranked by how close each launch is in meaning, then by votes. Prices were read from each product’s own site when checked and can change. Refine with your own description →