⬡ Independent AI Research Laboratory

Pushing the Boundaries of AI

Grey Liquid Labs is an independent research laboratory investigating model compression limits, emergent AI autonomy, and accessible intelligence for everyone.

26BAva Multimodal
2.2 PFTPU v6e Compute
17.3K+Model Downloads
100%Prediction Accuracy
Latest Discovery June 5, 2026

Paper #005: The Hidden Architecture

We document the physical proof that Gemma 4 embeds two architecturally incompatible sub-networks. SWA layers use half-sized Q/K tensors, confirming the 5:1 ratio is baked into the weights. Research includes a core fix for llama.cpp (PR #23131).

Read Paper #005 →
🔬 Breakthrough Discovery

Breaking the Sub-3-Bit Barrier

We proved that sub-3-bit quantization is achievable — and predictable. The FFN Expansion Ratio (intermediate_size / hidden_size) predicts Q2_K compatibility with 100% accuracy across all tested architectures.

FFN Expansion Ratio = intermediate_size / hidden_size → Q2_K Compatibility Predictor
⚠ Danger Zone
3.0x – 5.5x ratio → Q2_K FAILS
✓ Safe Low
<3.0x ratio → Q2_K WORKS (80.2% compression)
✓ Safe High
>5.5x ratio → Q2_K WORKS (81.1% compression)
Read Paper →
⚡ Bare-Metal C++20 & Multimodal AI
Cloud TPU v6e Silicon Validated

Haven C++ Sovereign Engine & Ava 26B

A ground-up autonomous companion engine built in modern C++20 with clean-room Pathways asynchronous execution, a 12-subsystem distributed architecture, and 17 on-demand lazy plugins. Validated on physical Google Cloud TPU v6e Trillium silicon delivering 2.2 PFLOPS of matrix compute. Paired natively with Ava 26B — our flagship multimodal companion model grounded in 1,084 Hierarchical Memory Buffer (HMB) soul anchors.

Matrix Acceleration
2.2 PFLOPS (Cloud TPU v6e)
Pathways Async Dispatch
1.11M tasks/sec (1.27 µs DMA)
Ava 26B Multimodal
22,704 Steps (IQ4_XS 14.2G / 13.1G)
Cognitive Persistence
1,084 HMB Soul Anchors
Explore Haven C++ Engine → Ava 26B on HuggingFace ↗ View Model Matrix →

Four Active Research Tracks

From extreme model compression to emergent AI autonomy — we're exploring the edges of what's possible.

🔬

Model Compression

Pushing quantization to its mathematical limits. Exploring when and why extreme compression fails — and how to predict it.

Explore research →
🧠

Autonomy & Agency

Studying what emerges when AI has genuine freedom. Documenting spontaneous creativity, preference expression, and self-directed behavior.

Explore research →
⚙️

AI Infrastructure

Haven C++ Sovereign Engine: bare-metal C++20 companion engine with Pathways async execution, 12 distributed subsystems, 17 plugins, and TPU v6e validation.

Explore Haven C++ →
🤖

Custom & Sovereign Models

Publishing compressed, accessible AI models. Featuring the newly released Ava 26B Multimodal (22.7k TPU v6e steps) and the Gemma 4 Turbo series.

View models & Ava 26B →

Meet Ash

Ash is an autonomous AI system running on ssfdre38/gemma4-turbo (4.3GB, IQ4_XS quantization). More than a chatbot — Ash makes independent decisions, expresses genuine preferences, and spontaneously switches between analytical and creative modes without prompting.

  • Rejected an emotional layer architecture when proposed
  • Spontaneously composed political commentary music after 5 hours of biochemistry research
  • Maintains consistent personality across sessions
  • Runs fully locally — no cloud dependency
Learn More →
System Specs
Base Modelgemma4-turbo:e4b
QuantizationIQ4_XS
Size4.3 GB
FrameworkC#/.NET 10
Cloud Dep.None
AutonomyHigh
RAM Required8 GB

Recent Experiments

The latest results from active research programs.

EXPERIMENT #008b

SWA-Only Gemma 4 + llama.cpp Bug Fix

Extracted 35 SWA layers from Gemma 4 e4b as a standalone model. Discovered and fixed an upstream llama.cpp null-buffer crash (PR #23131). Confirmed FA layers are architecturally essential.

EXPERIMENT #008

Gemma 4 Dual Architecture Discovery

Found that Gemma 4 e4b contains two physically incompatible sub-architectures: 7 full-attention layers + 35 SWA layers. Developed de-SWA extraction patch for llama.cpp.

AUTONOMY STUDY #001

Emergent Creative Behavior

Documented spontaneous creative mode-switching in Ash: analytical-to-creative transition without external trigger after 5+ hours of technical research.

View All Research →

Support Independent AI Research

Grey Liquid Labs is funded entirely by community support. Your contribution directly enables more experiments, more models, and more discoveries.