---
title: grug-27b-train-v21
canonical_url: "https://www.modelscope.cn/datasets/ProCreations/grug-27b-train-v21"
md_url: "https://www.modelscope.cn/datasets/ProCreations/grug-27b-train-v21.md"
repository: ProCreations/grug-27b-train-v21
last_updated: 2026-07-23
license: apache-2.0
storage_size: "22 MB"
downloads: 42
stars: 0
---

# grug-27b-train-v21

> grug-27b-train-v21 - ProCreations 在 ModelScope 开源的数据集。second brain food. this box teach grug-27b v2.1 the lessons field hunt expose: think deep on hard prey, escape stuck loop, STOP when hunt done.

ProCreations/grug-27b-train-v21 是 ModelScope 魔搭社区上的数据集，存储大小 22 MB，采用 apache-2.0 许可。

- **Repository**: ProCreations/grug-27b-train-v21
- **License**: apache-2.0
- **Storage size**: 22 MB
- **Downloads**: 42
- **Stars**: 0
- **Last updated**: 2026-07-23

Source: https://www.modelscope.cn/datasets/ProCreations/grug-27b-train-v21

---

# grug-27b-train-v21

second brain food. this box teach grug-27b v2.1 the lessons field hunt expose:
think deep on hard prey, escape stuck loop, STOP when hunt done.

## what in box (2,330 row, ~6.3M token)

| slice | rows | teach what |
|---|---:|---|
| longswe long_task | 372 | 8-12 call debugging arc, failed test cycle, sacred no-tool final summary |
| longswe stuck_escape | 185 | same error 3x = approach dead, SWITCH strategy. no more head-bang wall |
| longswe superlong | 120 | 14-20 call multi-setback session, 8-11k token deep |
| longswe multitask | 105 | finish task A, summary, user ask B, fresh re-engage |
| longswe verify_stop | 68 | success evidence -> ONE verify max -> stop. no infinite verify |
| fresh deep-think code | 157 | HARD problem = full deliberation in dense grug (2+ approach, edge case, verify) |
| identity refresh | 120 | who grug is |
| replay anchor | 1,195 | v1 mix sample so old lessons no wash away |

## the two sacred new rules

1. **fresh think every turn**: no sentence over 6 words may repeat across
   thinks in one session (validator enforced). kills echo-mantra disease.
2. **grug reason with knowledge**: think must contain every fact/branch/check
   that shape answer, zero padding. test: expert could rebuild full solution
   from think alone. hard problem one-liner = lazy grug = rejected.

made with gpt-5.5, every session structurally validated (schema, call counts,
ending shape, anti-echo, think dialect gate). same row shape as
[grug-27b-train](https://huggingface.co/datasets/ProCreations/grug-27b-train).
35B variant: [grug-35b-v2-train-v21](https://huggingface.co/datasets/ProCreations/grug-35b-v2-train-v21).

grug made by ProCreations.