feat: Protium-1.0 Objective-Gated Cognitive Architecture overhaul

README.md

@@ -1,157 +1,142 @@
 ---
-license: mit
 language:
 - en
 - zh
+license: apache-2.0
+library_name: transformers
 tags:
 - chat-template
-- jinja
-- hypersynapse-sas
-- cognitive-architecture
 - reasoning
-- multi-agent
-- swarm
-- deepseek
-- qwen
-- glm
-- llama
-- mistral
-- gemma
-- universal
-pipeline_tag: text-generation
+- cot
+- protium
+- solstice-ai
+- objective-gated
+- zero-waste
+- anti-mig25
+- gguf
+- ollama
+- llama.cpp
+- vllm
 ---
 
 <div align="center">
 
-# 🧠 HyperSynapse-SAS (SelfAgentSwarm) Chat Templates
-### *The Unified 10-Level Cognitive Escalation & Autonomous Multi-Agent Swarm Framework*
+# ⚡ Protium-1.0: Objective-Gated Cognitive Architecture
 
-[![Solstice AI](https://img.shields.io/badge/Solstice--AI-HyperSynapse--SAS-blueviolet?style=for-the-badge&logo=openai)](https://huggingface.co/Solstice-AI)
-[![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg?style=for-the-badge)](https://opensource.org/licenses/MIT)
-[![Model-Agnostic](https://img.shields.io/badge/Architecture-Model--Agnostic-success?style=for-the-badge)](https://huggingface.co/Solstice-AI/HyperSynapse-SAS-Chat-Templates)
-[![Levels](https://img.shields.io/badge/Cognitive%20Levels-0%20to%209%20Solo-orange?style=for-the-badge)](https://huggingface.co/Solstice-AI/HyperSynapse-SAS-Chat-Templates)
-[![Swarm](https://img.shields.io/badge/Swarm-20--Agent%20%2B%20Council-red?style=for-the-badge)](https://huggingface.co/Solstice-AI/HyperSynapse-SAS-Chat-Templates)
+[![Solstice AI](https://img.shields.io/badge/Solstice--AI-Protium--1.0-blueviolet?style=for-the-badge&logo=openai)](https://huggingface.co/Solstice-AI)
+[![Model-Agnostic](https://img.shields.io/badge/Architecture-Model--Agnostic-success?style=for-the-badge)](https://huggingface.co/Solstice-AI/Protium-1.0)
+[![Levels](https://img.shields.io/badge/Cognitive%20Spectrum-L0%20to%20L9%20Solo-orange?style=for-the-badge)](https://huggingface.co/Solstice-AI/Protium-1.0)
+[![Swarm](https://img.shields.io/badge/Swarm-20--Agent%20%2B%20Council-red?style=for-the-badge)](https://huggingface.co/Solstice-AI/Protium-1.0)
+
+**Zero Neutrons. Zero Dead Weight. Pure Thrust.**
 
 </div>
 
 ---
 
-## 🌟 Overview
+## 🔬 What is Protium-1.0?
 
-**HyperSynapse-SAS** (HyperSynapse Self-Agent Swarm) is a high-performance cognitive chat template framework that equips any Large Language Model with:
-1. **10-Level Solo Cognitive Escalation (`Level 0 Mortal` to `Level 9 Oracle`)**: Dynamically scale reasoning depth, token runway, and proof rigor per prompt without triggering multi-persona roleplay clutter.
-2. **Decoupled Multi-Agent Swarm Modes (`[swarm]` & `[deep-swarm]`)**: On-demand simulated 20-agent divergent swarms and adversarial consensus councils for massive, open-ended problem spaces.
-3. **Model-Agnostic / Universal Template**: A plug-and-play ChatML Jinja template compatible with **any modern LLM**, plus dedicated, fine-tuned templates for all major model families.
+**Protium-1.0** is Solstice-AI's high-efficiency, objective-gated cognitive reasoning architecture. Derived from **Protium** ($^1	ext{H}$) — the lightest, purest isotope of hydrogen consisting of one proton and zero neutrons — Protium-1.0 eliminates all cognitive ballast, rhetorical hesitation, and circular overthinking loops.
 
----
+Unlike legacy reasoning templates that impose arbitrary token floors (e.g. DavidAU's 5,000-token minimums), hardcoded pseudo-schemas, or unconstrained self-conversational babble, Protium-1.0 operates on a **forward-only, objective-gated protocol**.
 
-## 📁 Available Templates
+### Core Architecture Tenets
 
-| Template | File | Target Architectures & Families |
-| :--- | :--- | :--- |
-| **Universal (Model-Agnostic)** | [`templates/universal.jinja`](templates/universal.jinja) | Standard ChatML, vLLM, SGLang, llama.cpp, Ollama, LM Studio, Any LLM |
-| **DeepSeek** | [`templates/deepseek.jinja`](templates/deepseek.jinja) | DeepSeek-V3, DeepSeek-V4, DeepSeek-V4.1 Flash, DSML Tool Calling |
-| **Qwen** | [`templates/qwen.jinja`](templates/qwen.jinja) | Qwen 2.5, Qwen 3, Qwen 3.5, Qwen 3.8, Qwopus, QwQ |
-| **GLM** | [`templates/glm.jinja`](templates/glm.jinja) | GLM-4, GLM-5.3, GLM Flash |
-| **Llama** | [`templates/llama3.jinja`](templates/llama3.jinja) | Meta Llama 3, Llama 3.1, Llama 3.2, Llama 3.3, Llama 4 |
-| **Mistral** | [`templates/mistral.jinja`](templates/mistral.jinja) | Mistral 7B/12B, Mixtral 8x7B/8x22B, Codestral, Pixtral |
-| **Gemma** | [`templates/gemma.jinja`](templates/gemma.jinja) | Google Gemma 2, Gemma 3, Gemma 4 |
+1. **Token 1 Target Anchor**: The model isolates the exact termination condition at token 1. It anchors where "done and verified" lies before solving.
+2. **Forward-Only Deduction**: Thinking advances directly toward the target. Internal conversational monologues (*"Let's see..."*, *"Hmm..."*, *"Wait, let me rethink..."*) are strictly eliminated.
+3. **Objective-Gated Early Exit**: The moment the target condition is satisfied and verified, `<think>` terminates immediately. Zero stall, zero padding, zero manufactured doubt.
+4. **Anti-MiG-25 / Delta-Only Correction**: If a contradiction or constraint violation occurs, the model isolates and repairs only the specific broken delta. It never spirals into circular loops or complete restarts.
+5. **Absolute Operational Aura**: Clean, decisive, unhedged output delivery with maximum information density and zero conversational preamble.
 
 ---
 
-## 🎯 Cognitive Trigger Spectrum
-
-Simply prepend a tag to your query or pass `--level X` in your system instructions:
-
-### Solo Cognitive Escalation (Pure Analytical Focus)
+## 🎚️ The Cognitive Spectrum
+
+Protium-1.0 provides continuous elastic reasoning calibrated as typical output shares (never rigid token quotas):
+
+| Level | Identifier | Archetype | Target Share | Ideal Workload |
+|---|---|---|---|---|
+| **[L0]** | `mortal` / `regular` | Instant Execution | 0% (Thinking OFF) | Pure conversational chat, classification, summarization |
+| **[L1]** | `hermes` | Rapid Instinct | < 10% | Instant sanity check, syntax validation, arithmetic verify |
+| **[L2]** | `apollo` | Disciplined Deduction | 10% – 25% | Linear deduction, rule checking, constraint mapping |
+| **[L3]** | `artemis` | Boundary & Failure Audit | 20% – 40% | Edge cases, null checks, concurrency hazard scanning |
+| **[L4]** | `athena` | Strategic Architecture | 30% – 50% | Multi-step refactors, API design, trade-off optimization |
+| **[L5]** | `prometheus` | Defensive Engineering | 40% – 60% | Fault tolerance, scaling bottlenecks, lifecycle resilience |
+| **[L6]** | `solstice` | Deep Systemic Derivation | 50% – 70% | Dual-hypothesis synthesis, axiomatic proof, stress-testing |
+| **[L7]** | `hyperion` | First-Principles *(Default)* | 60% – 85% | Comprehensive first-principles decomposition |
+| **[L8]** | `einstein` | Theoretical Mastery | Majority | Theoretical unification, invariant reduction, deep algorithms |
+| **[L9]** | `oracle` | Proof-Grade Authority | Unbounded | Exhaustive formal proofs, zero-placeholder implementations |
+| **Swarm** | `swarm` | 20-Perspective Swarm | Elastic | Breakthrough ideation, divergent multi-perspective synthesis |
+| **Council**| `council` | Adversarial Research Panel | Elastic | Multi-specialist consensus, formal verification, audit |
 
-| Tag / Level | Persona / Archetype | Budget | Cognitive Modality |
-| :--- | :--- | :--- | :--- |
-| **`[Level 0]`** | **Mortal** | 0 tokens | **Instant Response**: CoT/thinking completely disabled. Rapid, direct execution with zero discursive fluff. |
-| **`[Level 1]`** | **Hermes** | ~256 tokens | **Rapid Instinct**: Single-path heuristic sanity check before emitting answer. |
-| **`[Level 2]`** | **Apollo** | ~512 tokens | **Crisp Logic**: Explicit step-by-step reasoning with edge-case validation. |
-| **`[Level 3]`** | **Artemis** | ~1,024 tokens | **Boundary Hunter**: Focuses on edge cases, off-by-one errors, and null states. |
-| **`[Level 4]`** | **Athena** | ~2,048 tokens | **Systemic Balance**: Second-order effects, system architecture, modular design. |
-| **`[Level 5]`** | **Prometheus** | ~4,096 tokens | **Defensive Engineering**: Fault tolerance, production bottlenecks, failure mode audit. |
-| **`[Level 6]`** | **Solstice** | ~8,192 tokens | **Deep Derivation**: Exhaustive theoretical derivation and mathematical mechanics. |
-| **`[Level 7]`** | **Hyperion (Default)** | ~16,384 tokens | **Foundational Synthesis**: First-principles breakdown from ground axioms. |
-| **`[Level 8]`** | **Einstein** | ~32,768 tokens | **Mastery & Invariants**: Theoretical proofs, deep algorithmic complexity analysis. |
-| **`[Level 9]`** | **Oracle** | Unbounded | **Maximal Rigor**: Exhaustive search runway, proof-grade verification, definitive closure. |
-
-### Autonomous Multi-Agent Swarm Modes
+---
 
-| Trigger | Mode Name | Protocol & Behavior |
-| :--- | :--- | :--- |
-| **`[swarm]`** | **20-Agent Divergent Swarm** | Activates a simulated 20-perspective emergent swarm (Analytical Logician, Creative Maverick, Systems Engineer, Red-Teamer, Optimizer, Critic, Synthesizer) with divergent exploration followed by structured synthesis. |
-| **`[deep-swarm]`** | **Adversarial Research Council** | Deploys a multi-round debate tournament pitting 3–5 specialized domain specialists against each other to falsify weak assumptions and converge on proof-grade deliverables. |
+## 🚀 Quickstart & Runtime Integration
 
----
+### 1. llama.cpp / LM Studio
+Pass any family template via the `--chat-template-file` flag:
+```bash
+llama-cli -m model.gguf --chat-template-file templates/qwen.jinja -p "[L4] Refactor this async worker"
+```
 
-## 🚀 Quick Start
+### 2. Ollama Modelfile
+```dockerfile
+FROM ./model.gguf
+TEMPLATE """{{ read_file "templates/universal.jinja" }}"""
+PARAMETER stop "<|im_start|>"
+PARAMETER stop "<|im_end|>"
+```
 
-### 1. Using the Universal Template in Python (`transformers`)
+### 3. Transformers (Python)
 ```python
 from transformers import AutoTokenizer
 
-tokenizer = AutoTokenizer.from_pretrained("Qwen/Qwen2.5-7B-Instruct")
-
-# Load the universal HyperSynapse-SAS template
-with open("templates/universal.jinja") as f:
+tokenizer = AutoTokenizer.from_pretrained("Qwen/Qwen2.5-32B-Instruct")
+with open("templates/qwen.jinja") as f:
     tokenizer.chat_template = f.read()
 
-messages = [
-    {"role": "user", "content": "[Level 3] Explain quantum decoherence rigorously."}
-]
-
-prompt = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
-print(prompt)
+# Invoke with runtime kwargs:
+prompt = tokenizer.apply_chat_template(
+    [{"role": "user", "content": "Prove that every planar graph is 5-colorable."}],
+    reasoning_effort="oracle",
+    tokenize=False,
+    add_generation_prompt=True
+)
 ```
 
-### 2. Apply Directly to Any Hugging Face Hub Repo
-Use our built-in CLI applier to update any repository or local folder with one command:
-```bash
-# Push to Hugging Face Model Hub
-python apply_template.py --template deepseek --model-id Solstice-AI/My-Model --push
+---
 
-# Or apply locally to a weights directory
-python apply_template.py --template universal --local-dir ./my_local_model
-```
+## 🎯 Steering Controls
 
-### 3. Using in `vLLM` or `SGLang`
-Pass `--chat-template templates/universal.jinja` directly at startup:
-```bash
-vllm serve meta-llama/Llama-3.1-8B-Instruct \
-  --chat-template templates/universal.jinja \
-  --port 8000
-```
+Protium-1.0 supports instant mid-session steering via inline tags or runtime kwargs:
 
-### 4. Using in `llama.cpp`
-```bash
-llama-cli \
-  --model my_model.gguf \
-  --chat-template-file templates/universal.jinja \
-  -p "<|im_start|>user\n[swarm] Architect an ultra-low latency event bus.<|im_end|>\n<|im_start|>assistant\n"
-```
-
-### 5. Ollama Modelfile
-```dockerfile
-FROM ./my_model.gguf
-TEMPLATE """{{ if .System }}<|im_start|>system
-{{ .System }}<|im_end|>
-{{ end }}{{ range .Messages }}<|im_start|>{{ .Role }}
-{{ .Content }}<|im_end|>
-{{ end }}<|im_start|>assistant
-"""
-PARAMETER stop "<|im_end|>"
-```
+- **Inline Steering**: Place tags anywhere in your prompt:
+  - `[L2]` or `{REASON:apollo}`: Sets Apollo level.
+  - `[L8]` or `{REASON:einstein}`: Sets Einstein theoretical level.
+  - `[regular]` or `[L0]`: Turns off thinking for pure vanilla response.
+  - `{REASON:i8}`: Silent execution (applies L8 cognitive rigor internally without emitting `<think>`).
+- **Runtime Kwargs**:
+  - `reasoning_effort="hermes"|"apollo"|"artemis"|"athena"|"prometheus"|"solstice"|"hyperion"|"einstein"|"oracle"|"swarm"|"council"`
+  - `thinking_level="10"` through `"100"` (continuous spectrum)
+  - `regular=True` or `mode="vanilla"` (zero overhead)
 
 ---
 
-## 📄 License
+## 📂 Model Family Templates
+
+| Family | Template File | Native Tokens |
+|---|---|---|
+| **Universal / ChatML** | `templates/universal.jinja` | `<\|im_start\|>`, `<\|im_end\|>` |
+| **Qwen 2.5 / 3.8** | `templates/qwen.jinja` | `<\|im_start\|>`, `<\|im_end\|>`, Vision |
+| **DeepSeek Native** | `templates/deepseek.jinja` | `<｜User｜>`, `<｜Assistant｜>`, DSML |
+| **Llama 3 / 4** | `templates/llama3.jinja` | `<\|start_header_id\|>`, `<\|eot_id\|>` |
+| **Mistral / Mixtral** | `templates/mistral.jinja` | `[INST]`, `[/INST]` |
+| **Google Gemma 2 / 3** | `templates/gemma.jinja` | `<start_of_turn>`, `<end_of_turn>` |
+| **Zhipu GLM 4 / 5** | `templates/glm.jinja` | `[gMASK]<sop>`, `<\|user\|>`, `<\|assistant\|>` |
+| **Vanilla Standalone** | `templates/regular_*.jinja` | Zero-overhead clean templates |
 
-Released under the permissive **MIT License**. Free for commercial and personal use across any open-weight model or inference stack.
+---
 
 <div align="center">
-<b>Developed by Solstice-AI</b><br>
-<i>Unshackled Cognition • Multi-Agent Collective Intelligence • Scalable Autonomous Systems</i>
+Built with precision by <b>Solstice AI</b>.
 </div>