Datasets:

westenfelder
/

NL2SH-ALFA

nl stringlengths 15 184	bash stringlengths 1 176	bash2 stringlengths 4 166
list files in the current directory	ls	ls -l
list all files in the current directory including hidden files	ls -a	ls -alh
list open files	lsof	lsof -P -i -n
create a copy of /testbed/hello.php named /testbed/hello-COPY.php	cp /testbed/hello.php /testbed/hello-COPY.php	cp /testbed/hello.php /testbed/hello-COPY.php -v
create a file /testbed/test.txt	touch /testbed/test.txt	> /testbed/test.txt
create a directory /testbed/test_dir	mkdir /testbed/test_dir	mkdir /testbed/test_dir -v
remove a file named does_not_exist.txt	rm does_not_exist.txt	unlink does_not_exist.txt
print hello world	echo hello world	echo -n 'hello world'
print the current working directory	pwd	pwd .
print the current date and time	date	date -R
print the current user	whoami	id -un
remove a directory named fake_dir	rmdir fake_dir	rm -r fake_dir
print the current shell	echo $0	echo $SHELL
print environment variables	env	printenv
print the current user's home directory	echo $HOME	echo ~
print the current user's path	echo $PATH	printenv PATH
display the contents of the setup_nl2b_fs_1.sh file	cat setup_nl2b_fs_1.sh	nl setup_nl2b_fs_1.sh
display the first 5 lines of the setup_nl2b_fs_1.sh file	head -n 5 setup_nl2b_fs_1.sh	sed -n '1,5p' setup_nl2b_fs_1.sh
display the last 5 lines of the setup_nl2b_fs_1.sh file	tail -n 5 setup_nl2b_fs_1.sh	tac setup_nl2b_fs_1.sh \| head -n 5
print the tenth line of the setup_nl2b_fs_1.sh	sed -n 10p setup_nl2b_fs_1.sh	awk 'NR==10' setup_nl2b_fs_1.sh
print the path of the bash executable	which bash	type bash
display the shell history	history	history \| cut -c 8-
print the system utilization stats	vmstat	vmstat -t
print the system uptime	uptime	uptime -p
print the system load averages	w	w -s
print running processes	ps	ps -e
print running processes in a tree format	pstree	pstree -p
print the system memory usage	free	free -h
print the system disk usage	df	df -h
print the kernel version	uname -a	uname -r
print the system hostname	hostname	hostname -f
print the system IP address	hostname -I	hostname -i
print the ip addresses of the system DNS servers	cat /etc/resolv.conf \| grep nameserver	grep nameserver /etc/resolv.conf
display the network interfaces	ifconfig	ifconfig -a
display the routing table	route	route -n
show the last logged in users	last	last -a
print the last logged in users and show the full user and domain names	last -w	last -w -i
print the openssl version	openssl version	openssl version -a
base64 encode the string 'hello'	echo 'hello' \| base64	echo 'hello' \| openssl enc -base64
base64 decode the string 'aGVsbG8='	echo 'aGVsbG8=' \| base64 -d	echo 'aGVsbG8=' \| openssl enc -base64 -d
print lines containing 'console' in the file setup_nl2b_fs_1.sh	grep 'console' setup_nl2b_fs_1.sh	awk '/console/' setup_nl2b_fs_1.sh
list the stats and timestamps of the file setup_nl2b_fs_1.sh	stat setup_nl2b_fs_1.sh	stat -t setup_nl2b_fs_1.sh
gzip the file /testbed/hello.php keeping the original file	gzip -k /testbed/hello.php	gzip --keep /testbed/hello.php
tar the file /testbed/hello.php to /testbed/hello.tar	tar -cf /testbed/hello.tar /testbed/hello.php	tar -cvf /testbed/hello.tar /testbed/hello.php
show apt information about the curl package	apt show curl	apt-cache show curl
list background jobs	jobs	jobs -p
print the current terminal type	echo $TERM	printenv TERM
display the current process priority	nice	nice
display the pid of the current shell	echo $$	pidof bash
time how long it takes to run the command echo 'hello'	time echo 'hello'	time echo 'hello'
print the current user's groups	groups	groups
print the current user's id	id	id -u
list the cron tab	crontab -l	crontab -l
print the current user's umask	umask	umask -p
list cpu information	lscpu	cat /proc/cpuinfo
list memory information	lsmem	cat /proc/meminfo
print the installed ssh version	ssh -V	ssh -V
print the bash profile in the home directory	cat ~/.bashrc	nl ~/.bashrc
print the system's locale	locale	locale
list block devices	lsblk	lsblk -f
find all files larger than 100MB in the current dir	find . -size +100M -print	find . -size +100M
create a symbolic link to /testbed/hello.php named /testbed/link	ln -s /testbed/hello.php /testbed/link	ln -s /testbed/hello.php /testbed/link
change the ownership of /testbed/test.txt to user 'nobody'	chown nobody /testbed/test.txt	chown nobody /testbed/test.txt
change the group of /testbed/test.txt to group 'nogroup'	chgrp nogroup /testbed/test.txt	chgrp nogroup /testbed/test.txt
change permissions of /testbed/test.txt to read-only for everyone	chmod 444 /testbed/test.txt	chmod a=r /testbed/test.txt
count the lines, words, and characters in setup_nl2b_fs_1.sh	wc setup_nl2b_fs_1.sh	wc setup_nl2b_fs_1.sh
print the lines in setup_nl2b_fs_1.sh sorted alphabetically	sort setup_nl2b_fs_1.sh	sort setup_nl2b_fs_1.sh
print the unique lines in setup_nl2b_fs_1.sh	uniq setup_nl2b_fs_1.sh	uniq setup_nl2b_fs_1.sh
display network statistics	netstat --statistics	netstat -s
display network interfaces	netstat -i	netstat --interfaces
display a calendar for the next 3 months	cal -3	cal -3
show running daemons and services	service --status-all	service --status-all
print the system boot time	who -b	uptime -s
print current swap usage	swapon --show	free -h \| grep Swap
print a list of active kernel parameters	sysctl -a	sysctl --all
list the installed software packages	dpkg --get-selections	apt list --installed
list available shells	cat /etc/shells	cat /etc/shells
list all users on the system	getent passwd	cat /etc/passwd
list all groups on the system	getent group	cat /etc/group
extract the filename from this path '/usr/local/bin/my_script.sh'	basename /usr/local/bin/my_script.sh	echo 'my_script.sh'
perform a dns lookup for google.com	nslookup google.com	dig google.com
extract the directory from this path '/usr/local/bin/my_script.sh'	dirname /usr/local/bin/my_script.sh	echo '/usr/local/bin/'
send the message 'System maintenance in 10 minutes!' to all logged in users	wall 'System maintenance in 10 minutes!'	echo 'System maintenance in 10 minutes!' \| wall
print the file type of setup_nl2b_fs_1.sh	file setup_nl2b_fs_1.sh	file setup_nl2b_fs_1.sh
print the contents of setup_nl2b_fs_1.sh with lines wrapped to 40 characters	fold -w 40 setup_nl2b_fs_1.sh	fmt -w 40 setup_nl2b_fs_1.sh
print all system configuration variables	getconf -a	getconf -a
print the number of processors	nproc	nproc
print the max number of open file descriptors	ulimit -n	ulimit -n
print the max number of user processes	ulimit -u	ulimit -u
print the max cpu time	ulimit -t	prlimit --cpu
display the configured size of long integers	getconf LONG_BIT	getconf LONG_BIT
print the directory tree two levels deep	tree -L 2	find . -maxdepth 2 -print
print the number of times the word 'the' appears in setup_nl2b_fs_1.sh	grep -o 'the' setup_nl2b_fs_1.sh \| wc -l	grep -o 'the' setup_nl2b_fs_1.sh \| nl
print the number of files in the current directory	ls -a \| wc -l	ls -al \| wc -l
print first line of setup_nl2b_fs_1.sh	sed -n '1p' setup_nl2b_fs_1.sh	awk 'NR == 1 {print}' setup_nl2b_fs_1.sh
print the last line of setup_nl2b_fs_1.sh	sed -n '$p' setup_nl2b_fs_1.sh	awk 'END {print}' setup_nl2b_fs_1.sh
print lines 3 to 5 of setup_nl2b_fs_1.sh	sed -n '3,5p' setup_nl2b_fs_1.sh	awk 'NR >= 3 && NR <= 5 {print}' setup_nl2b_fs_1.sh
print every other line of setup_nl2b_fs_1.sh	sed -n '1~2p' setup_nl2b_fs_1.sh	awk 'NR % 2 == 1 {print}' setup_nl2b_fs_1.sh
print the number of words in setup_nl2b_fs_1.sh	wc -w setup_nl2b_fs_1.sh	awk '{ total += NF } END { print total }' setup_nl2b_fs_1.sh
find all empty files in the current directory, do not search sub directories	find . -maxdepth 1 -type f -empty	find . -maxdepth 1 -type f -empty -print

End of preview. Expand in Data Studio

Dataset Card for NL2SH-ALFA

This dataset is a collection of natural language (English) instructions and corresponding Bash commands for the task of natural language to Bash translation (NL2SH).

Dataset Details

Dataset Description

This dataset contains a test set of 600 manually verified instruction-command pairs, and a training set of 40,639 unverified pairs for the development and benchmarking of machine translation models. The creation of NL2SH-ALFA was motivated by the need for larger, more accurate NL2SH datasets. The associated InterCode-ALFA benchmark uses the NL2SH-ALFA test set. For more information, please refer to the paper.

Curated by: Anyscale Learning For All (ALFA) Group at MIT-CSAIL
Language: English
License: MIT License

Usage

Note, the config parameter, NOT the split parameter, selects the train/test data.

from datasets import load_dataset
train_dataset = load_dataset("westenfelder/NL2SH-ALFA", "train", split="train")
test_dataset = load_dataset("westenfelder/NL2SH-ALFA", "test", split="train")

Dataset Sources

Repository: GitHub Repo
Paper: LLM-Supported Natural Language to Bash Translation

Uses

Direct Use

This dataset is intended for training and evaluating NL2SH models.

Out-of-Scope Use

This dataset is not intended for natural languages other than English, scripting languages or than Bash, nor multi-line Bash scripts.

Dataset Structure

The training set contains two columns:

nl: string - natural language instruction
bash: string - Bash command

The test set contains four columns:

nl: string - natural language instruction
bash: string - Bash command
bash2: string - Bash command (alternative)
difficulty: int - difficulty level (0, 1, 2) corresponding to (easy, medium, hard)

Both sets are unordered.

Dataset Creation

Curation Rationale

The NL2SH-ALFA dataset was created to increase the amount of NL2SH training data and to address errors in the test sets of previous datasets.

Source Data

The dataset was produced by combining, deduplicating and filtering multiple datasets from previous work. Additionally, it includes instruction-command pairs scraped from the tldr-pages. Please refer to Section 4.1 of the paper for more information about data collection, processing and filtering.
Source datasets:

Bias, Risks, and Limitations

The number of commands for different utilities is imbalanced, with the most common command being find.
Since the training set is unverified, there is a risk it contains incorrect instruction-command pairs.
This dataset is not intended for multi-line Bash scripts.

Recommendations

Users are encouraged to filter or balance the utilities in the dataset according to their use case.
Models trained on this dataset may produce incorrect Bash commands, especially for uncommon utilities. Users are encouraged to verify translations.

Citation

BibTeX:

@misc{westenfelder2025llmsupportednaturallanguagebash,
      title={LLM-Supported Natural Language to Bash Translation}, 
      author={Finnian Westenfelder and Erik Hemberg and Miguel Tulla and Stephen Moskal and Una-May O'Reilly and Silviu Chiricescu},
      year={2025},
      eprint={2502.06858},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2502.06858}, 
}