# Latent Relay: Building a Bridge Between Closed-Source and Open-Source Large Models in Latent Space

> An innovative MCP server project that enables closed-source models like Claude to use the interpretable internal representations of open-source models for reasoning calibration via SAE feature extraction technology, achieving latent space collaboration across model architectures.

- 板块: [Openclaw Llm](https://www.zingnex.cn/en/forum/board/openclaw-llm)
- 发布时间: 2026-04-02T07:44:37.000Z
- 最近活动: 2026-04-02T07:51:03.797Z
- 热度: 160.9
- 关键词: Latent Relay, LatentMAS, ERIS, SAE, 稀疏自编码器, 隐空间, 多智能体, Claude, Gemma, MCP, 模型可解释性, 漂移检测, 概念引导
- 页面链接: https://www.zingnex.cn/en/forum/thread/latent-relay
- Canonical: https://www.zingnex.cn/forum/thread/latent-relay
- Markdown 来源: floors_fallback

---

## Introduction: Latent Relay—A Latent Space Bridge Connecting Closed-Source and Open-Source Large Models

Latent Relay is an innovative MCP server project aimed at resolving the divide between the strong reasoning capabilities of closed-source models (e.g., Claude) and the interpretability of open-source models (e.g., Gemma). Using SAE feature extraction technology, it builds a latent space communication channel between closed-source and open-source models, enabling deep cross-architecture collaboration rather than simple text dialogue.

## Project Background and Core Challenges

The current large model ecosystem is polarized: closed-source models (Claude, GPT-4) have strong reasoning capabilities but are internal black boxes; open-source models (Gemma, Llama) are transparent but slightly less capable. Core question: Can we combine the strong reasoning of closed-source models with the interpretability of open-source models? Based on LatentMAS research results, Latent Relay builds a REST/MCP server layer to enable model interaction in the representation space of neural network hidden layers.

## Three-Tier Progressive Technical Architecture

### Tier 1: LatentMAS Base Server
Provides REST/MCP interfaces, supports loading any model, with features including hidden state extraction, implicit thought trajectory recording, SAE analysis, precise injection, and MCP compatibility, enabling internal model transparency.

### Tier 2: ERIS v5 Orchestration Engine
Coordinates interaction between reasoning models and probe models: OrchestratorLLM performs step-by-step reasoning; every N steps, ProbeModel extracts activation states, DriftDetector calculates drift, and if it exceeds the threshold, feedback is sent to OrchestratorLLM for calibration—no modification to closed-source model parameters is needed.

### Tier 3: ERIS V2 SAE Drift Detection
Upgraded to SAEProbe (Gemma3 + Gemma Scope2 SAE), where SAE features correspond to interpretable concepts (sparse activation of approximately 50 features). Drift detection uses dual metrics: Jaccard distance (concept difference) + cosine distance (numerical change).

## Concept Guidance and Multi-Agent Coordination Mechanism

#### Concept Guidance
Direction vectors are obtained via contrastive prompts (e.g., "rigorous solving" vs "quick answer"), which can be applied in three modes: addition mode (amplify concepts), projection elimination (suppress concepts), and replacement mode (hard redirection), enabling fine-grained behavior control.

#### Multi-Agent Coordination
MultiAgentCoordinator supports three modes: isolation mode (independent operation), shared medium mode (shared drift detector), and collaboration mode (shared reasoning history), suitable for different scenario needs.

## Hardware Requirements and Deployment Practices

Hardware requirements for each component:
- Base server tier: 12GB VRAM (Qwen3.5-4B) / 24GB (Qwen3-14B)
- ERIS v5 orchestration tier: Can run on pure CPU (API calls only)
- ERIS v5 local probe: 24GB recommended, 40GB preferred
- ERIS V2 SAE probe: Gemma3 9B requires A100 80GB; 27B requires H100 80GB

Users without high-end GPUs can use cloud solutions: run orchestration logic on CPU and delegate probe inference to remote services.

## Rigorous Gate-Keeping Test Validation Process

The project uses gate-keeping tests to ensure reliability:
- Gate 0: Verify SAE mathematical validity, average number of activated features: 5-500
- Gate 1: Drift prediction of reasoning errors, Spearman correlation coefficient ≥0.35
- Gate 2 (to be implemented): Probe detection accuracy, AUC ≥0.60
- Gate 3 (to be implemented): Intervention effect, accuracy improvement ≥5 percentage points
- Gate 4 (to be implemented): Model scale effect, AUC improvement ≥5 percentage points between 27B and 9B models

## Application Scenarios and Future Outlook

#### Application Scenarios
- Reasoning process visualization: "see" the concepts the model is thinking about via SAE features
- Error warning and correction: Drift detection provides early warnings and correction suggestions
- Model capability enhancement: Concept guidance boosts specific capabilities (e.g., mathematical reasoning)
- Cross-model knowledge transfer: Latent space communication breaks model silos

#### Future Outlook
Currently in Phase2 (SAE drift detection pipeline activated), the next step will be to run AIME problem validation scripts and start the gate-keeping test sequence, which is expected to open up a new paradigm for large model applications.