---
license: mit
base_model:
- zai-org/GLM-4-32B-0414
- zai-org/GLM-Z1-32B-0414
tags:
- merge
- mergekit
- slerp
library_name: transformers
language:
- en
- zh
- ko
---

# GLM-4-32B-0414-Z1-SLERP

A **SLERP merge** of `GLM-4-32B-0414` (instruct) and `GLM-Z1-32B-0414` (reasoning) at **t=0.3** (~70% instruct / 30% reasoning). No Korean fine-tuning — this is an experimental base artifact, shared transparently.

## TL;DR (honest)
On **GSM8K** (5-shot, 300 samples) this merge scores **94.0%** vs **88.3%** for its instruct parent — a **+5.7pp gain (~2.4 sigma, statistically significant)**. It absorbs some of the reasoning parent's math skill while keeping the instruct parent's concise answering. This is the **only** benchmark run. Not a SOTA claim.

## Results
| Model | GSM8K (5-shot) | n | stderr |
|---|---|---|---|
| **This merge (SLERP t=0.3)** | **0.940** | 300 | +/-0.014 |
| GLM-4-32B-0414 (instruct parent) | 0.883 | 300 | +/-0.019 |
| GLM-Z1-32B-0414 (reasoning parent) | 0.43* | 100 | +/-0.05 |

*The reasoning parent's low score is an **evaluation artifact**: its long chain-of-thought is truncated at the default generation length — not its true ability.

## Recipe
```yaml
merge_method: slerp
base_model: zai-org/GLM-4-32B-0414
models:
  - model: zai-org/GLM-Z1-32B-0414
parameters:
  t: 0.3
dtype: bfloat16
```

Made with [mergekit](https://github.com/arcee-ai/mergekit). Note: task-vector methods (`dare_ties` / `ties` with a base) collapsed this instruct+reasoning pair into incoherent multilingual/code output; only **SLERP** (direct interpolation, no base subtraction) yielded a coherent model.

## Limitations
- Only GSM8K evaluated — no general, safety, or Korean benchmarks.
- Instruct parent diluted 30%; other capabilities may regress.
- No Korean fine-tuning (future work).
- Transparent reproducible reference, not a production/SOTA model.

## Attribution
GLM-4-32B-0414 / GLM-Z1-32B-0414 (c) Zhipu AI (Z.ai), MIT License. Merged with mergekit (Arcee AI).