Show HN: SWE-ContextBench – A Benchmark for Context Learning in Coding Agents

IreneAI1 pts0 comments

GitHub - jiayuanz3/SWEContextBench · GitHub

/" data-turbo-transient="true" />

Skip to content

Type / to search

Sign in<br>Sign upAppearance settings

You signed in with another tab or window. Reload to refresh your session.<br>You signed out in another tab or window. Reload to refresh your session.<br>You switched accounts on another tab or window. Reload to refresh your session.

Dismiss alert

{{ message }}

jiayuanz3

SWEContextBench

Public

Notifications<br>You must be signed in to change notification settings

Fork

Star<br>12

main

BranchesTags

Go to file

CodeOpen more actions menu

Folders and files<br>NameNameLast commit message<br>Last commit date<br>Latest commit

History<br>20 Commits<br>20 Commits

assets

assets

cases

cases

predictions

predictions

swebench_memory

swebench_memory

.gitignore

.gitignore

README.md

README.md

environment.yml

environment.yml

evaluation.sh

evaluation.sh

View all files

Repository files navigation

SWE Context Bench: A Benchmark for Context Learning in Coding

SWE-ContextBench is a benchmark created to evaluate how well programming agents, such as AI coding systems, can reuse past experience when solving new tasks . It is built on top of existing datasets including SWE-Bench Lite, SWE-Bench Multilingual, and SWE-Bench Verified. The dataset contains 1,100 base tasks along with 376 related tasks that are derived from real dependency and reference relationships among GitHub issues and pull requests. These tasks are organized in a way that groups together problems with shared context, enabling the study of how effectively an agent can transfer knowledge across similar tasks . The dataset spans 51 real-world GitHub repositories and covers 9 different programming languages.

SWE-ContextBench is introduced as part of the research paper titled SWE Context Bench: A Benchmark for Context Learning in Coding.

SWE-ContextBench dataset can be downloaded from Hugging Face.

Run Evaluation

We provide pre-built Docker images for the related tasks.

Step1: Put your predictions in the predictions folder, with naming {instance_id}_preds.json

{instance_id}_preds.json follow the same data format convention as the SWE-Bench series datasets:

"instance_id": {<br>"model_name_or_path": {model_name_or_path},<br>"instance_id": {instance_id},<br>"model_patch": {model_patch}

Step2: Run

# Run only on "lite" subset<br>./evaluation.sh {run_id} lite

# Run on full dataset<br>./evaluation.sh {run_id} full

, {run_id} can be any text, e.g., my_run_id

🚨 News

02-05-26. Code Uploaded and Dataset Released 👩‍💻

About<br>No description, website, or topics provided.<br>Resources<br>Readme<br>Activity<br>Stars<br>12 stars<br>Watchers<br>0 watching<br>Forks<br>1 fork<br>Report repository

Releases

Packages

Contributors

Languages

You can’t perform that action at this time.

context bench tasks evaluation dataset instance_id

Related Articles