# InternLM/Intern-Decision

> A multimodal decision model built on Qwen3.5 that turns a state, optional images, and typed questions (choice, score, yes/no) into decisions with probabilities. Ships with training, two inference backends, temperature calibration, and a browser demo.

- Magnitude: 4.5 out of 10 — Early signal
- Stars: 78 total · +11 stars today, ≈ 40 by evening
- Star trust: star growth looks organic
- Category: Language models · Language: Python · Created: 2026-09-24 · Last push: 2026-09-26
- GitHub: https://github.com/InternLM/Intern-Decision · Page: https://gitnova.dev/en/r/InternLM/Intern-Decision

## Useful for

- Run local inference of Intern-Decision-4B on an RTX 4090 to measure latency and accuracy
- Calibrate model probabilities on a custom dataset via temperature scaling
- Train the model on your own records using the runtime schema and collation from the repo

## Why it’s here

- 11 stars so far today, about 40 expected by the end of the day.
- The spike has held for 2 days in a row — not a one-off blip.
- The repository is 3 days old and already has 78 stars. With less than two weeks of history, there's no usual pace to compare the spike against yet.
- Top new repositories this week: #189.
- About 9 forks a day — people are taking the code.

## 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: 9
- Issues and pull requests: 0
- Watchers: 1
- Average over the last week: 27 per day
- Usual pace: too little history (under two weeks)
- Stars in the last hour (measured): 3

## Stars per day, last 8 days (oldest → newest, today is partial)

2026-09-20 … 2026-09-27: 0, 0, 0, 0, 0, 0, 67, 11

## Spotted in now

- Top new repositories this week: #189

## Similar by description

1. **Liuziyu77/Valen** — 4.3 · Early signal · Language models · Python · +7 stars today, ≈ 40 by evening
   A multimodal decision model built on a Qwen3.5-0.8B/2B backbone: it takes text, images and video with an instruction and returns probabilities over supplied candidates without generating answer tokens. The repo includes model code, data…
   Full card: https://gitnova.dev/en/r/Liuziyu77/Valen.md
2. **fstandhartinger/jevbench** — 0.9 · Cooling · Language models · Python · +0 stars today, ≈ 2 by evening
   A benchmark for Jev-class decision models: the model receives state and a rubric, returns a typed answer with probabilities. It scores accuracy, calibration, speed, and cost.
   Full card: https://gitnova.dev/en/r/fstandhartinger/jevbench.md
3. **nokia-applied-research/AnyJev** — 3.7 · Cooling · Language models · Python · +1 star today, ≈ 23 by evening
   Turns any LLM into a decision model with typed questions (choice, yes/no, score) that return real probabilities read from the next-token distribution, without training. L0 removes position and label-prior bias; L1 adds temperature…
   Full card: https://gitnova.dev/en/r/nokia-applied-research/AnyJev.md
4. **togethercomputer/tev1** — 2.2 · Cooling · Language models · Python · +4 stars today, ≈ 9 by evening
   Open dataset and fine-tuning recipe for Qwen3.5-4B that picks one of 2–24 options given context and a question. The repo ships data, build scripts, Together AI training and evaluation tooling.
   Full card: https://gitnova.dev/en/r/togethercomputer/tev1.md
5. **kshetrajna12/reflex** — 0.7 · Steady · Language models · Python · +0 stars today, ≈ 1 by evening
   An open decision model: given a state (text, JSON, or image) and typed questions, it returns calibrated probabilities over fixed answer options. Runs on top of Qwen3.5 and answers with numbers instead of free text.
   Full card: https://gitnova.dev/en/r/kshetrajna12/reflex.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 2026-09-27 10:05 UTC, updated every 30 minutes.