Active NEPs

This page lists all currently active and pending NeuroShard Enhancement Proposals.

Status Overview

Status	Count	Description
🟢 ACTIVE	0	Currently enforced
🟡 VOTING	0	Open for voting
🔵 SCHEDULED	0	Approved, waiting activation
⚪ DRAFT	0	In preparation

Updated automatically from the network.

---

🟢 Active NEPs

No NEPs are currently active. The protocol is running on base version 1.0.0.

When NEPs are activated, they will appear here with:

• Implementation date
• Economic impact summary
• Version bump details

---

🟡 Currently Voting

No proposals are currently in the voting phase.

When a proposal enters voting, you can:

1. Review the specification
2. Understand the economic impact
3. Cast your vote (if you have staked NEURO)

---

🔵 Scheduled for Activation

No proposals are currently scheduled.

Approved proposals appear here with:

• Activation block number
• Days until activation
• Upgrade instructions

---

⚪ Draft Proposals

No proposals are currently in draft phase.

Draft proposals are being prepared and discussed before formal voting.

---

Proposed NEPs (Not Yet Submitted)

The following improvements are being discussed by the community:

NEP-DRAFT-001: Multi-Token Prediction (MTP)

Type: Training
Status: Community Discussion

Summary: Enable models to predict multiple future tokens per forward pass, extracting 2-3x more training signal.

Economic Impact:

• Training efficiency: +100%
• Per-batch reward: No change
• Net effect: Faster model improvement

Discussion: Discord #governance

---

NEP-DRAFT-002: Multi-Head Latent Attention (MLA)

Type: Architecture
Status: Technical Review

Summary: Compress KV cache by 20x using low-rank projection, enabling longer context on consumer hardware.

Economic Impact:

• Memory requirement: -2GB
• Inference speed: +30%
• Net effect: More nodes can participate

Discussion: Discord #governance

---

NEP-DRAFT-003: Auxiliary-Loss-Free MoE Balancing

Type: Training
Status: Research

Summary: Dynamic bias terms for MoE expert routing without auxiliary losses.

Economic Impact:

• Training stability: Improved
• Convergence: Faster
• Net effect: Better model quality

Discussion: Discord #governance

---

How to Track Proposals

Subscribe to Updates

# CLI notifications
neuroshard governance watch --notify email

# Or via API
curl https://api.neuroshard.com/governance/subscribe \
  -d '{"email": "you@example.com"}'

Query Programmatically

from neuroshard.core.governance import NEPRegistry, NEPStatus

registry = NEPRegistry()

# Get all voting proposals
voting = registry.list_proposals(status=NEPStatus.VOTING)
for nep in voting:
    print(f"{nep.nep_id}: {nep.title}")
    print(f"  Approval: {nep.approval_threshold:.0%}")
    print(f"  Ends: {nep.voting_end}")

# Get active NEPs
active = registry.list_proposals(status=NEPStatus.ACTIVE)

Via Website

Visit neuroshard.com/governance for:

• Real-time proposal status
• Vote casting interface
• Historical NEP archive

---

Historical Archive

All past proposals (approved and rejected) are archived for transparency.

NEP	Title	Status	Date
None yet	Protocol launched with base version	—	—

This table will populate as proposals are processed.

---

Submit Your Own

Have an idea to improve NeuroShard?

1. Discuss on Discord #governance
2. Write a proposal following the proposal guide
3. Submit with 100+ NEURO staked
4. Engage with community feedback

See How to Create a Proposal for detailed instructions.