Update README.md

README.md

@@ -13,7 +13,7 @@ pinned: false
 
 **Independent Canadian Open Reasoning Research**
 
-GestaltLabs is the research home for the **Nemostein**, **Ornstein**, **Harmonic**, **CRUCIBLE**, **Acta**, **Talos**, and related model lines by [DJLougen](https://huggingface.co/DJLougen). We build, curate, and release open-weight artifacts for practical reasoning systems: multimodal assistants, tool-using agents, local GGUF deployments, and post-training experiments around capability preservation, refusal shaping, and high-signal synthetic data.
+GestaltLabs is the research home for the **Nemostein**, **Ornstein**,  **CRUCIBLE**, **Acta**, **NSC-ACE**, and related model lines by [DJLougen](https://huggingface.co/DJLougen). We build, curate, and release open-weight artifacts for practical reasoning systems: multimodal assistants, tool-using agents, local GGUF deployments, and post-training experiments around capability preservation, refusal shaping, and high-signal synthetic data.
 
 [Models](https://huggingface.co/GestaltLabs?sort_models=lastModified) · [Datasets](https://huggingface.co/datasets?author=DJLougen) · [DJLougen](https://huggingface.co/DJLougen)
 
@@ -21,67 +21,6 @@ GestaltLabs is the research home for the **Nemostein**, **Ornstein**, **Harmonic
 
 ---
 
-## Current Focus
-- **Nemostein** — Nemotron/Nemotron-Omni reasoning and tool-use fine-tunes, including GGUF releases for local inference
-- **Ornstein** — multimodal and MoE reasoning models (Qwen/Gemma-derived) optimized for local deployment and agentic workflows
-- **CRUCIBLE** — Conditional Reward-Utility Coreset by Iterative Bayesian Learnability Estimation, a private data pipeline for reasoning and agent-trace curation
-- **SABER** — spectral refusal-boundary editing with KLD drift measurement for capability-preserving behavior edits
-- **RYS** — a separate post-training method line for refusal/behavior shaping experiments
-- **Acta / Talos** — curated agentic tool-use and coding-assistant traces for SFT and evaluation
-- **Local inference** — GGUF, quantized, and deployment-friendly builds for llama.cpp and MLX workflows
-
-## Featured Model Lines
-
-### Nemostein
-Nemotron 3 Nano / Nemotron Omni reasoning line focused on chat reasoning, Hermes-style tool calls, and local deployment.
-
-**Nemostein-3-Nano-Omni-30b-a3b** — Nemotron 3 Nano Omni lineage fine-tune for reasoning and tool-use chat
-- [BF16 / Transformers](https://huggingface.co/GestaltLabs/Nemostein-3-Nano-Omni-30b-a3b) · [GGUF](https://huggingface.co/GestaltLabs/Nemostein-3-Nano-Omni-30b-a3b-GGUF)
-- GGUF variants: Q8_0, Q6_K, Q4_K_M, Q2_K
-- Follow-on Acta Hermes tool-call SFT and SABER/KLD tuning are in progress.
-
-### Ornstein
-Multimodal and MoE reasoning line across 27B–35B scales, shipped in formats for every common local-inference stack.
-
-**Ornstein-3.6-27B** — multimodal base (Qwen 3.6 27B)
-- [Base](https://huggingface.co/GestaltLabs/Ornstein-3.6-27B) · [GGUF](https://huggingface.co/GestaltLabs/Ornstein-3.6-27B-GGUF)
-
-**Ornstein-Hermes-3.6-27b** — Hermes-format function-calling fine-tune (multimodal, tool-use)
-- [Base](https://huggingface.co/GestaltLabs/Ornstein-Hermes-3.6-27b) · [GGUF](https://huggingface.co/GestaltLabs/Ornstein-Hermes-3.6-27b-GGUF)
-- MLX: [bf16](https://huggingface.co/GestaltLabs/Ornstein-Hermes-3.6-27b-MLX-16bit) · [8-bit](https://huggingface.co/GestaltLabs/Ornstein-Hermes-3.6-27b-MLX-8bit) · [6-bit](https://huggingface.co/GestaltLabs/Ornstein-Hermes-3.6-27b-MLX-6bit) · [4-bit](https://huggingface.co/GestaltLabs/Ornstein-Hermes-3.6-27b-MLX-4bit)
-
-**Ornstein3.6-35B-A3B** — text-only Mixture-of-Experts (34.7B total, ~3B active)
-- [Base](https://huggingface.co/DJLougen/Ornstein3.6-35B-A3B) · [GGUF](https://huggingface.co/DJLougen/Ornstein3.6-35B-A3B-GGUF)
-- MLX: [bf16](https://huggingface.co/DJLougen/Ornstein3.6-35B-A3B-MLX-16bit) · [8-bit](https://huggingface.co/DJLougen/Ornstein3.6-35B-A3B-MLX-8bit) · [6-bit](https://huggingface.co/DJLougen/Ornstein3.6-35B-A3B-MLX-6bit) · [4-bit](https://huggingface.co/DJLougen/Ornstein3.6-35B-A3B-MLX-4bit)
-
-**Ornstein-27B-v2** — earlier 27B revision
-- [GGUF](https://huggingface.co/DJLougen/Ornstein-27B-v2-GGUF)
-
-### CRUCIBLE
-[CRUCIBLE](https://github.com/DJLougen/crucible) is a private data pipeline: Conditional Reward-Utility Coreset by Iterative Bayesian Learnability Estimation. It is used for reasoning-oriented data selection, self-correction supervision, and agent-trace curation.
-
-### SABER
-SABER is a spectral refusal-boundary editing toolkit that tunes candidates against generated refusal rate and KLD drift to preserve useful capability.
-
-- [SABER code](https://github.com/DJLougen/SABER)
-- [DJLougen/gemma-4-E2B-it-saber](https://huggingface.co/DJLougen/gemma-4-E2B-it-saber)
-
-### RYS
-RYS is a distinct post-training method line for refusal and behavior shaping experiments, tracked separately from SABER.
-
-## Code & Agent Infrastructure
-- [SABER](https://github.com/DJLougen/SABER) — refusal-boundary and KLD-drift tuning toolkit for capability-preserving behavior edits
-- [RBForge](https://github.com/DJLougen/RBForge) — runtime tool creation for Ornstein, Hermes, and SABER agents, with forged tools persisted into RBMEM memory
-- [Rust-Brain](https://github.com/DJLougen/Rust-Brain) — Rust rbmem CLI and structured memory format for timestamp-protected sections, hierarchy, graph relations, Hermes integration, and Markdown sync
-- [MPKnet](https://github.com/DJLougen/MPKnet) — bio-inspired CNN modeling the M/P/K pathways of the lateral geniculate nucleus; competitive accuracy with **52× fewer parameters than ResNet18**, designed for edge/embedded deployment
-
-## Featured Datasets
-- [Ornstein Curated 100K](https://huggingface.co/datasets/DJLougen/ornstein-curated-100k) — curriculum-sorted multi-domain reasoning data
-- [Hermes Agent Traces Filtered](https://huggingface.co/datasets/DJLougen/hermes-agent-traces-filtered) — quality-filtered agent reasoning traces
-- [Harmonic Reasoning v1](https://huggingface.co/datasets/DJLougen/harmonic-reasoning-v1) — compact synthetic reasoning data for math, code, and self-correction
-- [WittgenSite](https://huggingface.co/datasets/DJLougen/wittgensite) — prompt consistency benchmark for AI coding agents
-- [Acta](https://huggingface.co/datasets/DJLougen/Acta) — curated agentic tool-use conversations, including Hermes-style tool-call traces
-- [Talos Scenarios](https://huggingface.co/datasets/DJLougen/Talos-Scenarios) — agentic task scenarios for synthetic trace generation
 
 ## Principles
 - Ship **open artifacts** that people can inspect, run, and adapt