Massimo Roberto Scamarcia PRO

mrs83

AI & ML interests

Natural Language Processing, Text Generation, Question Answering, Data Augmentation, Knowledge Transfer, Chain-of-Thought, ResearchOps, MLOps

Recent Activity

new activity about 2 hours ago

ethicalabs/Echo-DSRN-114M-Base:Echo-DSRN-114M-Base v0.1.1

updated a model about 24 hours ago

ethicalabs/Echo-DSRN-114M-v0.1.1

updated a model 2 days ago

ethicalabs/Echo-DSRN-114M

View all activity

Organizations

posted an update 6 days ago

Post

131

Echo-DSRN-114M: A Constant-Memory O(1) Semantic Compressor

While traditional models target general conversational reasoning, Echo-DSRN(N) is a specialized structural prototype.

It is a dual-state recurrent neural network engineered strictly for low-latency semantic compression and continuous text streaming with a permanent O(1) memory footprint.

⚙️ Echo-DSRN (114M Parameters) manages context via continuous structural compression:

- 8 Layers | 512 Hidden Dim
- Transformer Fast State + DSRN/GRU Recurrent Slow State + Surprise Gating

Initial pre-training on a single AMD Instinct MI300X, followed by localized refinement across AMD Radeon PRO GPUs and an AMD Ryzen AI Max+ 395 (Strix Halo).

🖥️ A Hugging Face Space showcasing the architecture is currently running on the free shared CPU tier.

- The Compressor: Ingest a long document and crush it into a fixed 2048-dimensional .npy state vector.
- Vector Similarity: Upload two compressed .npy states to instantly calculate cosine similarity for ultra-lightweight RAG pre-filtering.
- The CPU Streamer: Continuous, fluent text generation running on raw CPU compute.

⚠️ Disclaimer: This is a structural prototype. It has internalized formatting and conversational syntax, but it possesses zero world knowledge. It will confidently hallucinate. Use it for streaming transcription, style mimicry, and local semantic hashing, not for factual reasoning.

Try the CPU Demo: ethicalabs/Echo-DSRN-114M
Try the Model: ethicalabs/Echo-DSRN-114M

reactedto SeaWolf-AI's post with 🔥 about 1 month ago

Post

8182

🚀 Introducing MARL — Runtime Middleware That Reduces LLM Hallucination Without Fine-Tuning

Now available on PyPI · GitHub · ClawHub · HuggingFace
AI models sense they could be wrong, but they can't actually fix what's broken.

🤗 Live A/B test: VIDraft/MARL

We evaluated 9 SOTA models (GPT-5.2, Claude Opus 4.6, Gemini 3 Pro, etc.) across 1,800 assessments in FINAL Bench and found a 39.2%p gap between "recognizing potential errors (MA=0.694)" and "actually finding and fixing them (ER=0.302)."

MARL (Model-Agnostic Runtime Middleware for LLMs) was built to close this metacognitive gap. It decomposes a single LLM call into a 5-stage expert pipeline (Hypothesis → Solver → Auditor → Adversarial Verifier → Synthesizer), transforming "answer in one shot" into "think, doubt, correct, and rewrite."

No weight modification — works instantly with GPT-5.4, Claude, Gemini, Llama, or any OpenAI API-compatible LLM by changing one line: base_url. Ships with 9 domain-specific emergence engines (invention, pharma, genomics, chemistry, ecology, law, and more — 5,538 expert data items) activated by a simple tag like model="gpt-5.4::pharma".

pip install marl-middleware

MARL is also officially registered on ClawHub, the skill marketplace of OpenClaw — an AI agent platform with 260K+ developers and 3,200+ skills. It's the first middleware in the Reasoning Enhancement category. One command — clawhub install marl-middleware — gives your AI agent a metacognition upgrade.

📝 Technical deep dive: https://huggingface.co/blog/FINAL-Bench/marl-middleware
📦 PyPI: https://pypi.org/project/marl-middleware/
🐙 GitHub: https://github.com/Vidraft/MARL
🦀 ClawHub: https://clawhub.ai/Cutechicken99/marl-middleware

#MARL #LLM #Hallucination #Metacognition #MultiAgent #AIMiddleware #FINALBench #OpenClaw #ClawHub #PyPI #AGI #HuggingFace #ReasoningAI #SelfCorrection #GlassBoxAI

reactedto SeaWolf-AI's post with 🔥 about 1 month ago

Post

11124

🏟️ Smol AI WorldCup: A 4B Model Just Beat 8B — Here's the Data

We evaluated 18 small language models from 12 makers on 125 questions across 7 languages. The results challenge the assumption that bigger is always better.

Community Article: https://huggingface.co/blog/FINAL-Bench/smol-worldcup
Live Leaderboard: ginigen-ai/smol-worldcup
Dataset: ginigen-ai/smol-worldcup

What we found:

→ Gemma-3n-E4B (4B, 2GB RAM) outscores Qwen3-8B (8B, 5.5GB). Doubling parameters gained only 0.4 points. RAM cost: 2.75x more.

→ GPT-OSS-20B fits in 1.5GB yet matches Champions-league dense models requiring 8.5GB. MoE architecture is the edge AI game-changer.

→ Thinking models hurt structured output. DeepSeek-R1-7B scores 8.7 points below same-size Qwen3-8B and runs 2.7x slower.

→ A 1.3B model fabricates confident fake content 80% of the time when prompted with nonexistent entities. Qwen3 family hits 100% trap detection across all sizes.

→ Qwen3-1.7B (1.2GB) outscores Mistral-7B, Llama-3.1-8B, and DeepSeek-R1-14B. Latest architecture at 1.7B beats older architecture at 14B.

What makes this benchmark different?

Most benchmarks ask "how smart?" — we measure five axes simultaneously: Size, Honesty, Intelligence, Fast, Thrift (SHIFT). Our ranking metric WCS = sqrt(SHIFT x PIR_norm) rewards models that are both high-quality AND efficient. Smart but massive? Low rank. Tiny but poor? Also low.

Top 5 by WCS:
1. GPT-OSS-20B — WCS 82.6 — 1.5GB — Raspberry Pi tier
2. Gemma-3n-E4B — WCS 81.8 — 2.0GB — Smartphone tier
3. Llama-4-Scout — WCS 79.3 — 240 tok/s — Fastest model
4. Qwen3-4B — WCS 76.6 — 2.8GB — Smartphone tier
5. Qwen3-1.7B — WCS 76.1 — 1.2GB — IoT tier

Built in collaboration with the FINAL Bench research team. Interoperable with ALL Bench Leaderboard for full small-to-large model comparison.

Dataset is open under Apache 2.0 (125 questions, 7 languages). We welcome new model submissions.

1 reply

reactedto SeaWolf-AI's post with 🔥 about 2 months ago

Post

4314

FINAL Bench Released: The Real Bottleneck to AGI Is Self-Correction

We release FINAL Bench, the first benchmark for measuring functional metacognition in LLMs — the ability to detect and correct one's own reasoning errors. Every existing benchmark measures final-answer accuracy. None measures whether AI knows it is wrong.

Dataset: [FINAL-Bench/Metacognitive]( FINAL-Bench/Metacognitive) | 100 Tasks | 15 Domains | 8 TICOS Types | Apache 2.0

Leaderboard: FINAL-Bench/Leaderboard

Article: https://huggingface.co/blog/FINAL-Bench/metacognitive

Core Innovation

Our 5-axis rubric separates what no prior benchmark could: MA (Metacognitive Accuracy) — the ability to say "I might be wrong", and ER (Error Recovery) — the ability to actually fix it. This maps directly to the monitoring-control model of Nelson & Narens (1990) in cognitive psychology.

Three Findings Across 9 SOTA Models

We evaluated GPT-5.2, Claude Opus 4.6, Gemini 3 Pro, DeepSeek-V3.2, Kimi K2.5, and others across 100 expert-level tasks:

1. ER Dominance. 94.8% of MetaCog gain comes from Error Recovery alone. The bottleneck to AGI is not knowledge or reasoning — it is self-correction.

2. Declarative-Procedural Gap. All 9 models can verbalize uncertainty (MA = 0.694) but cannot act on it (ER = 0.302). They sound humble but fail to self-correct — the most dangerous AI safety profile.

3. Difficulty Effect. Harder tasks benefit dramatically more from metacognition (Pearson r = -0.777, p < 0.001).

from datasets import load_dataset
dataset = load_dataset("FINAL-Bench/Metacognitive", split="train")

Paper: FINAL Bench: Measuring Functional Metacognitive Reasoning in LLMs

FINAL Bench is the first tool to tell apart what AI truly knows from what it merely pretends to know.

6 replies

repliedto qgallouedec's post about 2 months ago

Thanks for sharing, we are using a similar recipe for our small models 👏

reactedto qgallouedec's post with 🔥 about 2 months ago

Post

2999

@CohereLabs just released 🌿 Tiny Aya: a fully open-source 3B parameter model that speaks 70+ languages 🌍! But there’s a catch:

Tiny Aya is just a language model. It doesn’t support tool calling, the key capability that turns frontier models into powerful *agents*.
So the real question is:

How hard is it to turn Tiny Aya into an agent?

Turns out… it’s simple, thanks to Hugging Face TRL.
We’re sharing a hands-on example showing how to train Tiny Aya to turn it into a tool-calling agent using TRL, unlocking what could become the first *massively multilingual open agent*.

Small model. Global reach. Agent capabilities.

👉 https://github.com/huggingface/trl/blob/main/examples/notebooks/sft_tool_calling.ipynb

1 reply

repliedto their post about 2 months ago

https://huggingface.co/ethicalabs/Kurtis-EON1/discussions/2#699760c44c2b8775356cb36c

reactedto kostakoff's post with 🚀👍 about 2 months ago

Post

3345

My home lab for AI models - llmlaba v1

After I began learning MLOps I realized that I needed some kind of home lab, there are a lot of GPUs that I need to learn how to set up and test.
So I spent some time to do a researching which platform I could buy or build.
My requirements ware:
- Limited budget
- Power supply 1 kW or higher
- Few PCIe slots to be able to install more than one gpu
- Zero maintenance cost, I don't want spend a lot of time or money to maintain lab hardware, except for the GPUs

I chose the Intel Mac Pro 7.1:
- Prices on eBay acceptable
- Excelent cooling
- 1.4 kW power supply
- 7 PCIe slots
- Zero maintenance: I don't need to do anything with the Mac Pro hardware; it just works
- Classic UEFI boot loader

It requires a bit of OS preparation:
1. Install Ubuntu 24.04 (it works with the general PC ISO image)
2. Set up T2 drivers

sudo apt install -y dkms linux-headers-$(uname -r) applesmc-t2 apple-bce lm-sensors

3. Install t2fanrd to manually manage fans (/etc/t2fand.conf) https://wiki.t2linux.org/guides/fan/
4. Fix PCIe BAR: add pci=realloc to GRUB_CMDLINE_LINUX_DEFAULT so the Linux kernel will properly initializes server GPUs without Graphics Output Protocol
5. Install NVIDIA GPU driver:

sudo apt install nvidia-driver-570

And it works!
I was able to run server-grade Nvidia Tesla P100 (required DIY air duct), and consumer Nvidia Titan X, Titan V, GTX 1080 cards on the old Mac Pro 7.1 - even three in parallel.

llmlaba

3 replies

repliedto their post about 2 months ago

For MLX-LM we can only use Apple MPS. Re GPUs, at this moment I only have a AMD Ryzen AI Max+ 395 and AMD Instinct MI300X for rent.

repliedto their post 2 months ago

@maxxafits00 federated learning is definitely the path forward, and it’s something we’ve already begun experimenting with using the flower.ai framework.

Regarding the release, we are currently in mid-training and prioritizing a rigorous "safety-first" pipeline.

We are conducting extensive evaluations on model plasticity, red-teaming for prompt injection, and most importantly, stress-testing for malicious use cases.

We want to ensure the model is robust before it hits the wild.

The current roadmap includes:

Completing the Knowledge Expansion phase.
A comprehensive DPO (Direct Preference Optimization) pass to align the "Kurtis" persona and reasoning capabilities.
Peer review and final validation.

A quick technical spoiler:

The base model pre-training is pure PyTorch and fully multi-GPU compatible. We are utilizing a Curriculum Learning strategy: starting with a small context length and gradually scaling up. This is paired with an enormous batch size and small data chunks.

repliedto their post 2 months ago

This comment has been hidden

repliedto their post 2 months ago

@maxxafits00 if you are on a budget, I suggest to start small. I unfortunately don't have enough compute to scale right now. To evaluate a pretraining or distillation framework (such as arcee-ai's distillkit), or a new model architecture, you can start from datasets such as TinyStories and move to FineWeb-EDU, cosmopedia, etc later.

Wait for the training and architecture to be stable and validated before moving to a bigger dataset/model. Also, a 7-8B parameters is probably too big for small scale pre-training experiments.

You should try to target 0.5B, max 3B, especially if you use consumer-grade hardware, or a single GPU for rent.

repliedto their post 2 months ago

interesting. Yes, as you noticed as well a few billions tokens aren't enough. SmolLM2 360M was trained on 4 trillion tokens.

but I am not sure how to explain those results on piqa and sciq:

uv run lm_eval --model hf   --model_args pretrained=models/Echo-DSRN-Small-Instruct-Kurtis,trust_remote_code=True,device_map="auto"   --tasks hellaswag,winogrande,piqa,sciq --output_path ./results_final

Tasks	Version	Filter	Metric		Value		Stderr
hellaswag	1	none	acc	↑	0.2927	±	0.0045
		none	acc_norm	↑	0.3199	±	0.0047
piqa	1	none	acc	↑	0.6230	±	0.0113
		none	acc_norm	↑	0.6202	±	0.0113
sciq	1	none	acc	↑	0.7380	±	0.0139
		none	acc_norm	↑	0.6480	±	0.0151
winogrande	1	none	acc	↑	0.5020	±	0.0141

I can share more details in this convo, but this probably uncharted territory for an hybrid RNN with 4 attention heads

repliedto their post 2 months ago

Now available at https://huggingface.co/spaces/ethicalabs/Echo-DSRN-Small-Next-Word-Prediction ... on the shared CPU HF resources it runs slow, but on my Macbook M4 and AMD Strix Halo is blazing fast. Memory footprint is low. I am now expanding to 1B using Net2Net and today I tested a SFT run (QLoRA, 4-bit, bf16) on consumer hardware with trl with apparently no catastrophic forgetting.

posted an update 2 months ago

Post

2356

In 2017, my RNNs were babbling. Today, they are hallucinating beautifully.

10 years ago, getting an LSTM to output coherent English was a struggle.
10 years later, after a "cure" based on FineWeb-EDU and a custom synthetic mix for causal conversation, the results are fascinating.

We trained this on ~10B tokens on a single AMD GPU (ROCm). It is not a Transformer: Echo-DSRN (400M) is a novel recurrent architecture inspired by Hymba, RWKV, and xLSTM, designed to challenge the "Attention is All You Need" monopoly on the Edge.

The ambitious goal is to build a small instruct model with RAG and tool usage capabilities ( ethicalabs/Kurtis-EON1)

📊 The Benchmarks (Size: 400M)

For a model this size (trained on <10B tokens), the specialized performance is surprising:

*SciQ*: 73.8% 🦄 (This rivals billion-parameter models in pure fact retrieval).
*PIQA*: 62.3% (Solid physical intuition for a sub-1B model).

The Reality Check:

HellaSwag (29.3%) and Winogrande (50.2%) show the limits of 400M parameters and 10B tokens training.

We are hitting the "Reasoning Wall" which confirms we need to scale to (hopefully) unlock deeper common sense. As you can see in the visualization (to be released soon on HF), the FineWeb-EDU bias is strong. The model is convinced it is in a classroom ("In this course, we explore...").

The Instruct Model is not ready yet and we are currently using curriculum learning to test model plasticity.

Source code and weights will not be released yet. This is not a fork or a fine-tune: the base model is built in-house at https://www.ethicalabs.ai/, with novel components that do not exist in current open libraries.

🤝 Call for Collaboration: I am looking for Peer Reviewers interested in recurrent/hybrid architectures. If you want to explore what lies beyond Transformers, let’s connect!

Training diary: ethicalabs/Kurtis-EON1

6 replies

reactedto Reality123b's post with 👍 3 months ago

Post

1107

We (DataMuncher-Labs) just made a new dataset for inducing mathematical and analytical reasoning in LLMs.

DataMuncher-Labs/UltraMath-Reasoning-Small

1 reply

reactedto meg's post with ➕ 6 months ago

Post

4189

🤖 Did you know your voice might be cloned without your consent from just *one sentence* of audio?
That's not great. So with @frimelle , we brainstormed a new idea for developers who want to curb malicious use: ✨The Voice Consent Gate.✨
Details, code, here: https://huggingface.co/blog/voice-consent-gate

3 replies

posted an update 6 months ago

Post

239

Have you ever had a real conversation with a local LLM?

Or even taken a VoIP (SIP) phone call with one?

Check out Kurtis E1: A Fully On-Device MLX Voice Agent.

- GitHub Repo: https://github.com/ethicalabs-ai/Kurtis-E1-MLX-Voice-Agent
- Demo Video 1: https://www.youtube.com/watch?v=k6BbF0262I8
- Demo Video 2: https://www.youtube.com/shorts/w-YSCyKTR78

The entire stack runs on-device, leveraging MLX-LM for Apple Silicon.

- Whisper for real-time STT
- WebRTC VAD for voice activity detection
- A custom-tuned Qwen #LLM (Kurtis E1) ethicalabs/Kurtis-E1.1-Qwen2.5-3B-Instruct
- Coqui XTTSv2 for natural speech

This showcases the power of local AI/ML. I am also actively developing the SIP #VoIP integration (now in testing). The goal? To let you take a phone call and talk directly with your private agent, even without a computer or internet connection.

While Kurtis isn't built for math/coding, it shows a valuable path forward for on-device workflows.

We are actively looking for partners and clients to build out these POCs into real-world use cases.

https://www.ethicalabs.ai/ isn't a startup. We are not looking for VCs, equity deals, or grants: we're an open-source project.

If you like the R&D, you can support the R&D directly: https://github.com/sponsors/ethicalabs-ai?frequency=one-time

posted an update 7 months ago

Post

355

hallbayes https://github.com/leochlon/hallbayes is an interesting project by Leon Chlon (Hassana Labs) for checking for hallucination risk before text generation and it uses a powerful approach to decide if an LLM is confident enough to answer (or not).

https://arxiv.org/html/2509.11208v1
Predictable Compression Failures: Why Language Models Actually Hallucinate (2509.11208)

I've just integrated the hallbayes library into my completionist (synthetic dataset generation CLI tool) project to do exactly that, adding a new quality control layer to synthetic data generation.

Ran a small test on 10 samples on google/boolq with a 4B Qwen Instruct model Qwen/Qwen3-4B-Instruct-2507.

The output dataset now contains a hallucination_info column, flagging each sample with detailed metrics. The inference server is LM Studio, running on a Macbook Air M4 16GB

Test w/ hallucination flags: ethicalabs/google-boolq-hallbayes-test-qwen3-4b-2507

Implementation MRs:
https://github.com/leochlon/hallbayes/pull/16
https://github.com/ethicalabs-ai/completionist/pull/11

Massimo Roberto Scamarcia PRO

AI & ML interests

Recent Activity

Organizations

mrs83's activity