File size: 3,112 Bytes
f6e7a63
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
---
license: cc-by-4.0
task_categories:
  - text-generation
language:
  - en
tags:
  - shell
  - command-correction
  - terminal
  - synthetic
pretty_name: "Mash: mangled shell"
size_categories:
  - 10K<n<100K
---

# mash — mangled shell (training dataset)

~42k pairs for training shell-command **correction** models: garbled or
natural-language input → the command the user meant. As of Aug 2026 this task
had no public dataset — everything nearby is NL→bash translation. Built for
[fash](https://github.com/lumpenspace/fash)'s local fixer model,
[mash](https://huggingface.co/lumpenspace/mash).

## Format

mlx-lm chat JSONL (`train`/`valid`/`test`):

```json
{"messages": [
  {"role": "system", "content": "Return the shell command the user most likely wants. Reply with the command only."},
  {"role": "user", "content": "gti pusj origin mian"},
  {"role": "assistant", "content": "git push origin main"}
]}
```

## Composition

Valid commands come from [tldr-pages](https://github.com/tldr-pages/tldr)
(CC-BY-4.0, ~30k canonical example invocations) and
[NL2Bash](https://github.com/TellinaTool/nl2bash) (MIT, ~12k expert
one-liners). Placeholders are filled from concrete scenarios, with descriptions
rewritten to match (so "open a specific file" becomes "open the settings file"
→ `joe settings.json`).

A share of rows (~11%) carry a **context header**`cwd:` plus a `files:`
directory listing — synthesized so exactly one plausible variant of the
referenced file exists; the answer must use the file that is actually there
("open the settings file" → `joe settings.toml` when the listing has
`settings.toml`, not `settings.json`). The rest are bare, so models trained on
this data work with or without context.

Three pair types:

- **typo pairs** — commands corrupted by a weighted taxonomy of realistic
  manglings: transposed/dropped/adjacent-QWERTY/doubled chars, `-`/`--`
  confusion, merged words, stripped quotes, missing or spurious `sudo`,
  copy-pasted `$ ` prompt prefixes, duplicated words, smart quotes/en-dashes,
  plus curated high-frequency head typos (`gerp`, `sl`, `suod`, `dokcer`, …).
- **request pairs** — example descriptions, verbatim and "casualized" into the
  terse phrasing people actually type.
- **identity pairs** (~5%) — correct→correct, so models learn not to
  over-correct. Kept deliberately small: at ~12% a 1.5B run learned to echo
  39% of genuinely garbled inputs.

Generated deterministically by
[`training/build_dataset.py`](https://github.com/lumpenspace/fash/blob/main/training/build_dataset.py)
(seed 7). No user data or shell history is included.

## Evaluation

Exact match on the held-out `test` split flatters models — it shares the
generator's bias. For real-world numbers use
[mash-bench](https://huggingface.co/datasets/lumpenspace/mash-bench), 203
scenarios extracted from human breakage (NoFAQ, thefuck).

## Attribution

Derived from [tldr-pages](https://github.com/tldr-pages/tldr) (CC-BY-4.0) and
[NL2Bash](https://github.com/TellinaTool/nl2bash) (MIT; Lin et al., LREC 2018).
Dataset licensed CC-BY-4.0 accordingly.