---
pretty_name: MorphBench (German)
language:
- de
license: cc-by-sa-4.0
tags:
- morphology
- tokenization
- evaluation
- german
dataset_info:
- config_name: task1_inflection
  features:
  - name: lemma
    dtype: string
  - name: feats
    dtype: string
  - name: target
    dtype: string
  - name: status
    dtype: string
  splits:
  - name: train
    num_bytes: 835801
    num_examples: 11982
  - name: dev
    num_bytes: 69889
    num_examples: 995
  - name: test
    num_bytes: 138820
    num_examples: 1989
  - name: test_rare
    num_bytes: 36938
    num_examples: 499
  - name: test_memorization
    num_bytes: 32659
    num_examples: 498
  download_size: 393231
  dataset_size: 1114107
- config_name: task2_deriv_segmentation
  features:
  - name: word
    dtype: string
  - name: segmentation
    dtype: string
  - name: status
    dtype: string
  splits:
  - name: train
    num_bytes: 583722
    num_examples: 10000
  - name: dev
    num_bytes: 59481
    num_examples: 1000
  - name: test_main
    num_bytes: 118920
    num_examples: 1966
  - name: test_memorization
    num_bytes: 28135
    num_examples: 500
  - name: test_oov
    num_bytes: 31843
    num_examples: 500
  download_size: 348072
  dataset_size: 822101
- config_name: task3_derivation
  features:
  - name: base
    dtype: string
  - name: affix
    dtype: string
  - name: target
    dtype: string
  - name: status
    dtype: string
  splits:
  - name: train
    num_bytes: 610293
    num_examples: 9935
  - name: dev
    num_bytes: 62389
    num_examples: 998
  - name: test_main
    num_bytes: 126360
    num_examples: 1992
  - name: test_memorization
    num_bytes: 29265
    num_examples: 494
  - name: test_oov
    num_bytes: 33410
    num_examples: 500
  download_size: 331468
  dataset_size: 861717
- config_name: task4_compound
  features:
  - name: word
    dtype: string
  - name: segmentation
    dtype: string
  - name: difficulty
    dtype: string
  - name: source
    dtype: string
  splits:
  - name: train
    num_bytes: 573352
    num_examples: 9982
  - name: dev
    num_bytes: 57319
    num_examples: 997
  - name: test_main
    num_bytes: 109111
    num_examples: 2000
  - name: test_rare
    num_bytes: 6426
    num_examples: 117
  - name: test_memorization
    num_bytes: 30049
    num_examples: 499
  - name: test_hardest
    num_bytes: 30053
    num_examples: 495
  download_size: 437099
  dataset_size: 806310
- config_name: task5_affix_function
  features:
  - name: word
    dtype: string
  - name: function
    dtype: string
  - name: freq
    dtype: int64
  - name: status
    dtype: string
  splits:
  - name: train
    num_bytes: 345423
    num_examples: 5462
  - name: dev
    num_bytes: 47312
    num_examples: 747
  - name: test
    num_bytes: 98358
    num_examples: 1552
  download_size: 108713
  dataset_size: 491093
- config_name: task6a_definition
  features:
  - name: word
    dtype: string
  - name: gloss
    dtype: string
  - name: base
    dtype: string
  - name: affix
    dtype: string
  - name: function
    dtype: string
  - name: derived_freq
    dtype: int64
  - name: base_exposure
    dtype: int64
  - name: status
    dtype: string
  splits:
  - name: train
    num_bytes: 1243576
    num_examples: 9131
  - name: dev
    num_bytes: 165878
    num_examples: 1251
  - name: test_main
    num_bytes: 42888
    num_examples: 327
  - name: test_main_oov
    num_bytes: 68787
    num_examples: 480
  - name: test_rare
    num_bytes: 51435
    num_examples: 379
  - name: test_memorization
    num_bytes: 144907
    num_examples: 1022
  download_size: 946934
  dataset_size: 1717471
- config_name: task6b_definition_mcq
  features:
  - name: word
    dtype: string
  - name: base
    dtype: string
  - name: affix
    dtype: string
  - name: function
    dtype: string
  - name: gold_gloss
    dtype: string
  - name: distractors
    sequence: string
  - name: pretrain_status
    dtype: string
  - name: derived_freq
    dtype: int64
  - name: base_exposure
    dtype: int64
  splits:
  - name: dev
    num_bytes: 580104
    num_examples: 1242
  - name: main
    num_bytes: 152581
    num_examples: 324
  - name: main_oov
    num_bytes: 224637
    num_examples: 473
  - name: rare
    num_bytes: 178322
    num_examples: 375
  - name: memorization
    num_bytes: 480506
    num_examples: 1013
  download_size: 767173
  dataset_size: 1616150
configs:
- config_name: task1_inflection
  data_files:
  - split: train
    path: task1_inflection/train-*
  - split: dev
    path: task1_inflection/dev-*
  - split: test
    path: task1_inflection/test-*
  - split: test_rare
    path: task1_inflection/test_rare-*
  - split: test_memorization
    path: task1_inflection/test_memorization-*
- config_name: task2_deriv_segmentation
  data_files:
  - split: train
    path: task2_deriv_segmentation/train-*
  - split: dev
    path: task2_deriv_segmentation/dev-*
  - split: test_main
    path: task2_deriv_segmentation/test_main-*
  - split: test_memorization
    path: task2_deriv_segmentation/test_memorization-*
  - split: test_oov
    path: task2_deriv_segmentation/test_oov-*
- config_name: task3_derivation
  data_files:
  - split: train
    path: task3_derivation/train-*
  - split: dev
    path: task3_derivation/dev-*
  - split: test_main
    path: task3_derivation/test_main-*
  - split: test_memorization
    path: task3_derivation/test_memorization-*
  - split: test_oov
    path: task3_derivation/test_oov-*
- config_name: task4_compound
  data_files:
  - split: train
    path: task4_compound/train-*
  - split: dev
    path: task4_compound/dev-*
  - split: test_main
    path: task4_compound/test_main-*
  - split: test_rare
    path: task4_compound/test_rare-*
  - split: test_memorization
    path: task4_compound/test_memorization-*
  - split: test_hardest
    path: task4_compound/test_hardest-*
- config_name: task5_affix_function
  data_files:
  - split: train
    path: task5_affix_function/train-*
  - split: dev
    path: task5_affix_function/dev-*
  - split: test
    path: task5_affix_function/test-*
- config_name: task6a_definition
  data_files:
  - split: train
    path: task6a_definition/train-*
  - split: dev
    path: task6a_definition/dev-*
  - split: test_main
    path: task6a_definition/test_main-*
  - split: test_main_oov
    path: task6a_definition/test_main_oov-*
  - split: test_rare
    path: task6a_definition/test_rare-*
  - split: test_memorization
    path: task6a_definition/test_memorization-*
- config_name: task6b_definition_mcq
  data_files:
  - split: dev
    path: task6b_definition_mcq/dev-*
  - split: main
    path: task6b_definition_mcq/main-*
  - split: main_oov
    path: task6b_definition_mcq/main_oov-*
  - split: rare
    path: task6b_definition_mcq/rare-*
  - split: memorization
    path: task6b_definition_mcq/memorization-*
- config_name: verb_cloze
  data_files:
  - split: train
    path: verb_cloze/train.jsonl
  - split: validation
    path: verb_cloze/validation.jsonl
  - split: test
    path: verb_cloze/test.jsonl
  - split: test_rare
    path: verb_cloze/test_rare.jsonl
---

# MorphBench — German Morphology Evaluation Suite

A benchmark for **morphology-aware tokenizers and language models** in German.
Eight numbered tasks probe inflection, segmentation, derivation, compounding, affix
semantics, and whole-word meaning. Each task is a **config**; difficulty/generalization
tiers are **splits**.

```python
from datasets import load_dataset
ds = load_dataset("yuanxin112/morphbench-de", "task3_derivation", split="test_main")
```

## Tasks

### `task1_inflection`
Produce the inflected form from a lemma + features.
**input** `interagieren` + `v;ind;pl;1;pst` → **output** `interagierten`
columns: `lemma, feats, target, status`

### `task2_deriv_segmentation` — *derivational segmentation*
Split a **derived** word into base + affix.
**input** `Dreher` → **output** `drehen|er`
columns: `word, segmentation, status`

### `task3_derivation`
Build the derived word from a base + affix.
**input** `Apotheke` + `pflichtig` → **output** `apothekenpflichtig`
columns: `base, affix, target, status`

### `task4_compound`
Split a compound into its parts.
**input** `Freiheitsheld` → **output** `Freiheit|Held`
columns: `word, segmentation, difficulty, source`

### `task5_affix_function`
Given a derived word, classify the **semantic function of its affix** (13 classes:
`negation, without, agent_person, make_become, …`).
**input** `stimmlos` → **output** `without`
columns: `word, function, freq, status`

### `task6a_definition`
Generate the dictionary gloss of a derived word.
**input** `Malocher` → **output** `Arbeitnehmer, der überwiegend körperlich hart … arbeitet`
columns: `word, gloss, base, affix, function, derived_freq, base_exposure, status`

### `task6b_definition_mcq`
Multiple-choice version of 6a: pick the correct gloss among hard distractors.
columns: `word, base, affix, function, gold_gloss, distractors, pretrain_status, derived_freq, base_exposure`

## Splits — how items are bucketed

Every test split is defined by the **pretraining exposure** of the item's parts
(frequency in the BabyLM-style German training corpus): *seen* = frequency above a
threshold, *unseen* = frequency 0. The headline idea everywhere is **the target is
UNSEEN while its base/lemma IS seen** (compositional generalization); `memorization`
and `rare`/`OOV`/`hardest` are the seen / harder controls. `train`, `dev` are
training / validation.

**`task1_inflection`** — by *(lemma seen?)* × *(this form seen?)*:

| split | condition |
|-------|-----------|
| `test` (main) | lemma seen, form **unseen** |
| `test_memorization` | lemma seen, form seen |
| `test_rare` | lemma **unseen**, form unseen |

**`task2_deriv_segmentation`, `task3_derivation`** — by *(base seen?)* × *(derived seen?)*, where *seen* = corpus freq ≥ 1:

| split | condition |
|-------|-----------|
| `test_main` | base seen, derived **unseen** (compositional generalization) |
| `test_memorization` | base seen, derived seen |
| `test_OOV` | base **unseen**, derived unseen |

**`task4_compound`** — by *(compound word seen?)* × *(all components seen?)*, where compound *seen* = freq ≥ 1 and a component is *seen* = freq ≥ 5:

| split | condition | what it tests |
|-------|-----------|---------------|
| `test_main` | compound **unseen**, all components seen | compositional generalization |
| `test_memorization` | compound seen, all components seen | rote recall |
| `test_rare` | compound seen, some component rare | mixed |
| `test_hardest` | compound **unseen**, some component unseen | hardest |

**`task6a_definition`, `task6b_definition_mcq`** — by the derived word's corpus frequency `derived_freq`, splitting the *unseen* case by `base_exposure`:

| split | condition |
|-------|-----------|
| `memorization` | `derived_freq ≥ 20` |
| `rare` | `derived_freq` 1–5 |
| `main` | `derived_freq = 0` and `base_exposure ≥ 300` (base familiar) |
| `main_oov` | `derived_freq = 0` and `base_exposure < 300` (base also OOV) |

**`task5_affix_function`** uses a plain `train` / `dev` / `test` split (no difficulty buckets). The `status` / `pretrain_cell` column on each row
records the exposure cell directly, e.g. `base_seen+derived_unseen`, `lem_seen+form_unseen`.

> Note: the German "seen" thresholds are lower than the English suite (derivation/
> segmentation use freq ≥ 1; definition uses `derived_freq ≥ 20`, `base_exposure ≥ 300`).

## Provenance

Built from German Wiktionary (via [kaikki.org](https://kaikki.org) / wiktextract) and UniMorph.
Companion lexical resource: [`yuanxin112/wiktionary-morph`](https://huggingface.co/datasets/yuanxin112/wiktionary-morph).
English counterpart: [`yuanxin112/morphbench-en`](https://huggingface.co/datasets/yuanxin112/morphbench-en).

## License

Derived from Wiktionary — [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/);
attribute Wiktionary and its contributors.

## `verb_cloze` — Contextual Verb-Inflection Cloze (added config)

Leave-one-out cloze: given a verb `lemma`, partial `feats` (one morphological
dimension withheld), and a natural sentence with the target verb replaced by a
blank marker, generate the inflected form. Splits are **lemma-disjoint**.
Sentences come from Universal Dependencies treebanks (a *derivative*; CC BY-SA
4.0 — verify per-treebank licenses, e.g. English ParTUT is CC BY-NC-SA).

```python
from datasets import load_dataset
ds = load_dataset("yuanxin112/morphbench-de", "verb_cloze")
```