# MUSE-AI-Lab/StateTrack

> StateTrack-Bench is an evaluation protocol for long-horizon dialogues that plants a persistent constraint, probes it at increasing depths, and uses a forked recall audit to separate forgetting from remembering-without-complying (State Integrity Collapse). It also ships a training-free mitigation, Structured State Anchoring (SSA).

- Magnitude: 5.2 out of 10 — Early signal
- Stars: 261 total · +76 stars measured 2026-10-30, 07:15–08:40 UTC, ≈ 189 by evening
- Star trust: star growth looks organic
- Category: Language models · Language: Python · License: MIT · Created: 2026-10-27 · Last push: 2026-10-29
- GitHub: https://github.com/MUSE-AI-Lab/StateTrack · Page: https://gitnova.dev/en/r/MUSE-AI-Lab/StateTrack

## Useful for

- Reproduce the paper's retention tables offline from cached outputs without an API key
- Run the core or augmented benchmark suite against a chosen model to measure constraint retention
- Apply Structured State Anchoring to reduce behavioral drift in long dialogues

## Why it’s here

- Star-counter measurements on 2026-10-30 (UTC), 07:15–08:40: 185 → 261 stars (+76). This is the change over that interval.
- Estimated end-of-day forecast: about +189 stars, using observed gains and the previous day.
- The repository is 3 days old and already has 261 stars. With less than two weeks of history, there's no usual pace to compare the spike against yet.
- Top new repositories this week: #80.

## Star trust

Star growth looks organic.

Star-trust labels are heuristics based on the repository’s behavior, not a check of every stargazer.

## Numbers

- Forks: 0
- Issues and pull requests: 6
- Watchers: 0
- Average over the last week: 90 per day
- Usual pace: too little history (under two weeks)
- Stars in the last hour (measured): 54

## Stars per day, last 4 days (oldest → newest, today is partial)

2026-10-27 … 2026-10-30: 0, 5, 164, 76

## Spotted in now

- Top new repositories this week: #80

## More in this category

1. **togg53192-cmd/jailbreaks** — 8.9 · Breakout · Language models · +685 stars measured 2026-10-30, 00:15–08:40 UTC, ≈ 1,923 by evening
   A collection of 31 system prompts, persona locks, and AGENTS.md files for 25+ models (Claude, Gemini, Grok, GLM, DeepSeek, Kimi, Qwen, Gemma, GPT-OSS). Intended for research on how instruction hierarchies and policy overrides affect model…
   Full card: https://gitnova.dev/en/r/togg53192-cmd/jailbreaks.md
2. **routerhubSI/openrouter-alternative** — 6.0 · Early signal · Language models · +0 stars measured 2026-10-30, 00:23–08:40 UTC, ≈ 58 by evening
   An AI model gateway offering one prepaid API key with OpenAI- and Anthropic-compatible interfaces, including separate uncensored and abliterated DeepSeek, GLM and Qwen entries. The repository is a migration guide from OpenRouter plus a…
   Full card: https://gitnova.dev/en/r/routerhubSI/openrouter-alternative.md
3. **OpenBMB/MiniCPM-OCR** — 5.9 · Early signal · Language models · TypeScript · +64 stars measured 2026-10-30, 00:16–08:40 UTC, ≈ 189 by evening
   MiniCPM-OCR is an open-source OCR model family for parsing Chinese and English documents and PDFs into Markdown or JSON, with a 2B model (Jev) that scores recognition quality per region.
   Full card: https://gitnova.dev/en/r/OpenBMB/MiniCPM-OCR.md
4. **Edge0-AI/Edge0** — 5.5 · Breakout · Language models · Python · +46 stars measured 2026-10-30, 00:16–08:40 UTC, ≈ 377 by evening
   An open-source streaming MoE inference framework: expert weights are offloaded from SSD on demand while a trained prerouter predicts routing ahead of time. Runs on Apple Silicon via MLX and ships with two ready-to-run model tiers (35B and…
   Full card: https://gitnova.dev/en/r/Edge0-AI/Edge0.md
5. **MUSE-AI-Lab/ReMCTS** — 5.2 · Early signal · Language models · Python · +54 stars measured 2026-10-30, 07:15–08:40 UTC, ≈ 161 by evening
   ReMCTS is a research framework that improves LLM code generation by searching over a tree of candidate programs. Each candidate is executed, and failures are turned into structured reflections that guide the search, outperforming direct…
   Full card: https://gitnova.dev/en/r/MUSE-AI-Lab/ReMCTS.md

---

Magnitude (0–10) measures how fast and how unusually interest in a repository is growing right now. It is not a quality score.
Days are UTC. “So far today” is a fact; “expected by the end of the day” is a forecast.
Summaries and use cases are written by an LLM (DeepSeek V4.1 Flash) from the README and may be inaccurate: verify specific claims (benchmarks, speed, hardware) in the repository itself.
Data as of 2026-10-30 08:51 UTC, updated every 30 minutes.