instance_id large_stringlengths 12 57 | messages listlengths 45 198 | family large_stringclasses 1
value | repo large_stringlengths 8 53 | language large_stringclasses 1
value | verified bool 1
class | resolved large_stringclasses 1
value | patch large_stringlengths 261 658k | model large_stringclasses 1
value | harness large_stringclasses 1
value | id_kind large_stringclasses 3
values | upstream_id large_stringlengths 12 57 | affine_source large_stringclasses 1
value | origin large_stringclasses 1
value | source_uid large_stringlengths 80 84 | submitted bool 2
classes | affine_fold_group large_stringclasses 0
values |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
0b01001001__spectree-64 | [{"role":"system","content":"You are a helpful assistant that can interact multiple times with a com(...TRUNCATED) | pr | 0b01001001__spectree | python | true | resolved | "diff --git a/setup.py b/setup.py\nindex 1b3cb64..4ef21e6 100644\n--- a/setup.py\n+++ b/setup.py\n@@(...TRUNCATED) | Qwen/Qwen3-Coder-480B-A35B-Instruct | OpenHands | pr | 0b01001001__spectree-64 | SWE-Lego | SWE-Lego | PrimeIntellect/SWE-Lego-Real-Data-Verified|data/resolved-00000-of-00003.parquet|405 | true | null |
12rambau__sepal_ui-747 | [{"role":"system","content":"You are a helpful assistant that can interact multiple times with a com(...TRUNCATED) | pr | 12rambau__sepal_ui | python | true | resolved | "diff --git a/sepal_ui/sepalwidgets/sepalwidget.py b/sepal_ui/sepalwidgets/sepalwidget.py\nindex 408(...TRUNCATED) | Qwen/Qwen3-Coder-480B-A35B-Instruct | OpenHands | pr | 12rambau__sepal_ui-747 | SWE-Lego | SWE-Lego | PrimeIntellect/SWE-Lego-Real-Data-Verified|data/resolved-00000-of-00003.parquet|941 | true | null |
12rambau__sepal_ui-758 | [{"role":"system","content":"You are a helpful assistant that can interact multiple times with a com(...TRUNCATED) | pr | 12rambau__sepal_ui | python | true | resolved | "diff --git a/sepal_ui/sepalwidgets/inputs.py b/sepal_ui/sepalwidgets/inputs.py\nindex 04e69553..0cb(...TRUNCATED) | Qwen/Qwen3-Coder-480B-A35B-Instruct | OpenHands | pr | 12rambau__sepal_ui-758 | SWE-Lego | SWE-Lego | PrimeIntellect/SWE-Lego-Real-Data-Verified|data/resolved-00001-of-00003.parquet|546 | true | null |
12rambau__sepal_ui-814 | [{"role":"system","content":"You are a helpful assistant that can interact multiple times with a com(...TRUNCATED) | pr | 12rambau__sepal_ui | python | true | resolved | "diff --git a/sepal_ui/sepalwidgets/alert.py b/sepal_ui/sepalwidgets/alert.py\nindex 19718f51..8dafa(...TRUNCATED) | Qwen/Qwen3-Coder-480B-A35B-Instruct | OpenHands | pr | 12rambau__sepal_ui-814 | SWE-Lego | SWE-Lego | PrimeIntellect/SWE-Lego-Real-Data-Verified|data/resolved-00000-of-00003.parquet|902 | true | null |
15five__scim2-filter-parser-20 | [{"role":"system","content":"You are a helpful assistant that can interact multiple times with a com(...TRUNCATED) | pr | 15five__scim2-filter-parser | python | true | resolved | "diff --git a/src/scim2_filter_parser/parser.py b/src/scim2_filter_parser/parser.py\nindex 516f65d..(...TRUNCATED) | Qwen/Qwen3-Coder-480B-A35B-Instruct | OpenHands | pr | 15five__scim2-filter-parser-20 | SWE-Lego | SWE-Lego | PrimeIntellect/SWE-Lego-Real-Data-Verified|data/resolved-00000-of-00003.parquet|389 | true | null |
2gis__k8s-handle-120 | [{"role":"system","content":"You are a helpful assistant that can interact multiple times with a com(...TRUNCATED) | pr | 2gis__k8s-handle | python | true | resolved | "diff --git a/k8s_handle/templating.py b/k8s_handle/templating.py\nindex 7f2d6b7..445a09e 100644\n--(...TRUNCATED) | Qwen/Qwen3-Coder-480B-A35B-Instruct | OpenHands | pr | 2gis__k8s-handle-120 | SWE-Lego | SWE-Lego | PrimeIntellect/SWE-Lego-Real-Data-Verified|data/resolved-00002-of-00003.parquet|1106 | true | null |
2gis__k8s-handle-73 | [{"role":"system","content":"You are a helpful assistant that can interact multiple times with a com(...TRUNCATED) | pr | 2gis__k8s-handle | python | true | resolved | "diff --git a/.dockerignore b/.dockerignore\nindex d3f33a2..fea1e67 100644\n--- a/.dockerignore\n+++(...TRUNCATED) | Qwen/Qwen3-Coder-480B-A35B-Instruct | OpenHands | pr | 2gis__k8s-handle-73 | SWE-Lego | SWE-Lego | PrimeIntellect/SWE-Lego-Real-Data-Verified|data/resolved-00002-of-00003.parquet|1247 | true | null |
3yourmind__django-migration-linter-186 | [{"role":"system","content":"You are a helpful assistant that can interact multiple times with a com(...TRUNCATED) | pr | 3yourmind__django-migration-linter | python | true | resolved | "diff --git a/CHANGELOG.md b/CHANGELOG.md\nindex d1ec8e5..15fefc0 100644\n--- a/CHANGELOG.md\n+++ b/(...TRUNCATED) | Qwen/Qwen3-Coder-480B-A35B-Instruct | OpenHands | albedo:open-swe-traces | 3YOURMIND__django-migration-linter-186 | SWE-Lego | SWE-Lego | PrimeIntellect/SWE-Lego-Real-Data-Verified|data/resolved-00000-of-00003.parquet|1429 | true | null |
3yourmind__django-migration-linter-258 | [{"role":"system","content":"You are a helpful assistant that can interact multiple times with a com(...TRUNCATED) | pr | 3yourmind__django-migration-linter | python | true | resolved | "diff --git a/CHANGELOG.md b/CHANGELOG.md\nindex 3069d91..beafd65 100644\n--- a/CHANGELOG.md\n+++ b/(...TRUNCATED) | Qwen/Qwen3-Coder-480B-A35B-Instruct | OpenHands | albedo:open-swe-traces | 3YOURMIND__django-migration-linter-258 | SWE-Lego | SWE-Lego | PrimeIntellect/SWE-Lego-Real-Data-Verified|data/resolved-00001-of-00003.parquet|532 | true | null |
42DIGITAL__bqtools-11 | [{"role":"system","content":"You are a helpful assistant that can interact multiple times with a com(...TRUNCATED) | pr | 42DIGITAL__bqtools | python | true | resolved | "diff --git a/fourtytwo/bqtools/__init__.py b/fourtytwo/bqtools/__init__.py\nindex 9542242..136e18c (...TRUNCATED) | Qwen/Qwen3-Coder-480B-A35B-Instruct | OpenHands | pr | 42DIGITAL__bqtools-11 | SWE-Lego | SWE-Lego | PrimeIntellect/SWE-Lego-Real-Data-Verified|data/resolved-00001-of-00003.parquet|1142 | true | null |
End of preview. Expand in Data Studio
affine-openhands
4,231 coding-agent trajectories on 4,231 real pull-request
tasks, converted to the format of albedo's mini-coder source (mini-swe-agent: one ```bash command per assistant
turn, <returncode>/<output> observations, mini-coder's system prompt and task template).
- Origin: PrimeIntellect/SWE-Lego-Real-Data-Verified (OpenHands transcripts by Qwen3-Coder-480B-A35B-Instruct; both its
resolvedanddroppedfiles). - Machine: SWE-Lego's OpenHands runs (interactive terminal:
lsin columns,git logwith branch labels, the repository on its own branch). Albedo tells machines apart by source name, so each kind of machine is its own dataset. - Repository root: rewritten to
/testbedin every command and output.
How it was built
From affine-dedup (Dendrite):
- Task-level dedup: tasks matching one of our benchmarks were dropped. Runs on tasks albedo already has were kept, under albedo's own task id when that id points at the same checkout the run had.
- Trajectory-level dedup: no run is identical to another or to any run albedo already has.
- Conversion: every tool call became one bash command with what that command prints; edits became a
python3heredoc that replaces the exact text; harness wrappers, to-do and memory tools were removed. - Solution leaks dropped: runs that got the fix from outside their own work (SWE-smith's
Bug Patchhistory, the task's own PR ingit log --all, fetching upstream code over the network). - Benchmark leaks dropped with albedo's
scripts/benchmark_leaks.py. - Validated against mini-coder's format with albedo's own format checks.
Across all seven datasets: 10,456 runs exported, 9,958 converted, 2,288 dropped as solution leaks, 113 as benchmark leaks, 7,557 published.
Contents
| Resolved | Rows |
|---|---|
| resolved | 4,231 |
| Harness | Rows |
|---|---|
| OpenHands | 4,231 |
| Task pool | Rows |
|---|---|
| SWE-Lego | 4,231 |
| Language | Rows |
|---|---|
| python | 4,231 |
Columns
| Column | Meaning |
|---|---|
instance_id |
task id, in the style albedo's repo-context service resolves to a repository and commit |
messages |
the trajectory: system prompt, task, then assistant turns (THOUGHT + one ```bash command) and <returncode>/<output> observations, mini-coder's format |
family |
bug family: pr for real pull requests; SWE-smith's lm, combine, mechanical otherwise |
repo |
owner__repo |
language |
repository language |
verified |
true when the run is known to have resolved the task |
resolved |
resolved, not resolved or unknown (Affine's own verdict where it gave one) |
patch |
the task's reference patch where the upstream record has one, else empty |
model |
model that produced the run |
harness |
agent harness the run was recorded in |
id_kind |
where instance_id comes from: albedo's own id for the task, Open-SWE's, or the task pool's |
upstream_id |
the task id in the upstream record |
affine_source |
task pool (swesmith, scaleswe, multiswe, swerebench_v2, r2e_gym, swelego, or SWE-Lego) |
origin |
Affine corpus or SWE-Lego |
source_uid |
where the raw run is: Affine chunk URL and line, or SWE-Lego file and row |
submitted |
whether the run ends with the submit command (false: cut by the step limit) |
affine_fold_group |
Affine's reason for keeping the run (completion, king_fail, ...), when it gave one |
License
Apache-2.0, as upstream SWE-Lego.
- Downloads last month
- 60