Hugging Face's logo

Datasets:
rootxhacker
/
patchbench

Dataset card Data Studio Files
xet
 Community
Dataset Preview
Duplicate
The full dataset viewer is not available (click to read why). Only showing a preview of the rows.
The dataset generation failed
Error code:   DatasetGenerationError
Exception:    CastError
Message:      Couldn't cast
id: string
repo: string
base_commit: string
status: string
verify_status: string
error: null
timing: struct<setup: double, pre_test_0: double, pre_test_1: double, post_test_0: double, post_test_1: doub (... 18 chars omitted)
  child 0, setup: double
  child 1, pre_test_0: double
  child 2, pre_test_1: double
  child 3, post_test_0: double
  child 4, post_test_1: double
  child 5, total: double
test_cmds: list<item: string>
  child 0, item: string
rebuild_cmds: list<item: null>
  child 0, item: null
print_cmds: list<item: null>
  child 0, item: null
f2p_pre_ok: bool
pre_detail: struct<missing: int64, wrong: list<item: null>, missing_ids: list<item: null>, parsed_count: int64>
  child 0, missing: int64
  child 1, wrong: list<item: null>
      child 0, item: null
  child 2, missing_ids: list<item: null>
      child 0, item: null
  child 3, parsed_count: int64
f2p_post_ok: bool
p2p_post_ok: bool
post_detail: struct<missing: int64, wrong: list<item: null>, missing_ids: list<item: null>, parsed_count: int64>
  child 0, missing: int64
  child 1, wrong: list<item: null>
      child 0, item: null
  child 2, missing_ids: list<item: null>
      child 0, item: null
  child 3, parsed_count: int64
raw_tail: string
context_files: list<item: struct<path: string, content: string>>
  child 0, item: struct<path: string, content: string>
      child 0, path: string
      child 1, content: string
context_lines: int64
buggy_files: list<item: string>
  child 0, item: string
test_files: list<item: struct<path: string, content: string>>
  child 0, item: struct<path: string, content: string>
      child 0, path: string
      child 1, content: string
eval: struct<test_cmds: list<item: string>, print_cmds: list<item: null>, log_parser: string, rebuild_cmds (... 19 chars omitted)
  child 0, test_cmds: list<item: string>
      child 0, item: string
  child 1, print_cmds: list<item: null>
      child 0, item: null
  child 2, log_parser: string
  child 3, rebuild_cmds: list<item: null>
      child 0, item: null
origin: string
upstream_instance_id: string
source: string
issue_text: string
hints_text: string
log_parser: string
test_patch: string
fail_to_pass: list<item: string>
  child 0, item: string
language: string
docker_image: string
gold_patch: string
pass_to_pass: list<item: string>
  child 0, item: string
to
{'id': Value('string'), 'language': Value('string'), 'source': Value('string'), 'origin': Value('string'), 'upstream_instance_id': Value('string'), 'repo': Value('string'), 'base_commit': Value('string'), 'issue_text': Value('string'), 'hints_text': Value('string'), 'gold_patch': Value('string'), 'test_patch': Value('string'), 'fail_to_pass': List(Value('string')), 'pass_to_pass': List(Value('string')), 'test_cmds': List(Value('string')), 'log_parser': Value('string'), 'docker_image': Value('string'), 'context_files': Value('null')}
because column names don't match
Traceback:    Traceback (most recent call last):
                File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1827, in _prepare_split_single
                  for key, table in generator:
                                    ^^^^^^^^^
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 613, in wrapped
                  for item in generator(*args, **kwargs):
                              ~~~~~~~~~^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 343, in _generate_tables
                  self._cast_table(pa_table, json_field_paths=json_field_paths),
                  ~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 132, in _cast_table
                  pa_table = table_cast(pa_table, self.info.features.arrow_schema)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2378, in table_cast
                  return cast_table_to_schema(table, schema)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2306, in cast_table_to_schema
                  raise CastError(
                  ...<3 lines>...
                  )
              datasets.table.CastError: Couldn't cast
              id: string
              repo: string
              base_commit: string
              status: string
              verify_status: string
              error: null
              timing: struct<setup: double, pre_test_0: double, pre_test_1: double, post_test_0: double, post_test_1: doub (... 18 chars omitted)
                child 0, setup: double
                child 1, pre_test_0: double
                child 2, pre_test_1: double
                child 3, post_test_0: double
                child 4, post_test_1: double
                child 5, total: double
              test_cmds: list<item: string>
                child 0, item: string
              rebuild_cmds: list<item: null>
                child 0, item: null
              print_cmds: list<item: null>
                child 0, item: null
              f2p_pre_ok: bool
              pre_detail: struct<missing: int64, wrong: list<item: null>, missing_ids: list<item: null>, parsed_count: int64>
                child 0, missing: int64
                child 1, wrong: list<item: null>
                    child 0, item: null
                child 2, missing_ids: list<item: null>
                    child 0, item: null
                child 3, parsed_count: int64
              f2p_post_ok: bool
              p2p_post_ok: bool
              post_detail: struct<missing: int64, wrong: list<item: null>, missing_ids: list<item: null>, parsed_count: int64>
                child 0, missing: int64
                child 1, wrong: list<item: null>
                    child 0, item: null
                child 2, missing_ids: list<item: null>
                    child 0, item: null
                child 3, parsed_count: int64
              raw_tail: string
              context_files: list<item: struct<path: string, content: string>>
                child 0, item: struct<path: string, content: string>
                    child 0, path: string
                    child 1, content: string
              context_lines: int64
              buggy_files: list<item: string>
                child 0, item: string
              test_files: list<item: struct<path: string, content: string>>
                child 0, item: struct<path: string, content: string>
                    child 0, path: string
                    child 1, content: string
              eval: struct<test_cmds: list<item: string>, print_cmds: list<item: null>, log_parser: string, rebuild_cmds (... 19 chars omitted)
                child 0, test_cmds: list<item: string>
                    child 0, item: string
                child 1, print_cmds: list<item: null>
                    child 0, item: null
                child 2, log_parser: string
                child 3, rebuild_cmds: list<item: null>
                    child 0, item: null
              origin: string
              upstream_instance_id: string
              source: string
              issue_text: string
              hints_text: string
              log_parser: string
              test_patch: string
              fail_to_pass: list<item: string>
                child 0, item: string
              language: string
              docker_image: string
              gold_patch: string
              pass_to_pass: list<item: string>
                child 0, item: string
              to
              {'id': Value('string'), 'language': Value('string'), 'source': Value('string'), 'origin': Value('string'), 'upstream_instance_id': Value('string'), 'repo': Value('string'), 'base_commit': Value('string'), 'issue_text': Value('string'), 'hints_text': Value('string'), 'gold_patch': Value('string'), 'test_patch': Value('string'), 'fail_to_pass': List(Value('string')), 'pass_to_pass': List(Value('string')), 'test_cmds': List(Value('string')), 'log_parser': Value('string'), 'docker_image': Value('string'), 'context_files': Value('null')}
              because column names don't match
              
              The above exception was the direct cause of the following exception:
              
              Traceback (most recent call last):
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 1369, in compute_config_parquet_and_info_response
                  parquet_operations, partial, estimated_dataset_info = stream_convert_to_parquet(
                                                                        ~~~~~~~~~~~~~~~~~~~~~~~~~^
                      builder, max_dataset_size_bytes=max_dataset_size_bytes
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                  )
                  ^
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 948, in stream_convert_to_parquet
                  builder._prepare_split(split_generator=splits_generators[split], file_format="parquet")
                  ~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1694, in _prepare_split
                  for job_id, done, content in self._prepare_split_single(
                                               ~~~~~~~~~~~~~~~~~~~~~~~~~~^
                      gen_kwargs=gen_kwargs, job_id=job_id, **_prepare_split_args
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                  ):
                  ^
                File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1880, in _prepare_split_single
                  raise DatasetGenerationError("An error occurred while generating the dataset") from e
              datasets.exceptions.DatasetGenerationError: An error occurred while generating the dataset

Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.

id
string
language
string
source
string
origin
string
upstream_instance_id
string
repo
string
base_commit
string
issue_text
string
hints_text
string
gold_patch
string
test_patch
string
fail_to_pass
list
pass_to_pass
list
test_cmds
list
log_parser
string
docker_image
string
context_files
null
patchbench-go-real-0028
go
real
swe-bench-live-multilang
knative__eventing-8611
knative/eventing
5bcf8011acc8310ca0e469e507c1b4c757de4fff
"RequestReply: Create Data Plane\n<!-- \r\nAre you using Knative? If you do, we would love to know!\(...TRUNCATED)
/assign /triage accepted
"diff --git a/pkg/apis/eventing/v1alpha1/register.go b/pkg/apis/eventing/v1alpha1/register.go\nindex(...TRUNCATED)
"diff --git a/pkg/requestreply/ingress_handler_test.go b/pkg/requestreply/ingress_handler_test.go\nn(...TRUNCATED)
["knative.dev/eventing/pkg/requestreply/TestHandlerServeHttp","knative.dev/eventing/pkg/requestreply(...TRUNCATED)
["knative.dev/eventing/pkg/apis/config/TestGetBrokerConfig/Namespace_default_with_given_broker_class(...TRUNCATED)
[ "/usr/local/go/bin/go test -v ./... -json > reports/go-test-results.json" ]
"def parser(log: str) -> dict[str, str]:\n import json\n import re\n\n results: dict[str, s(...TRUNCATED)
starryzhang/sweb.eval.x86_64.knative_1776_eventing-8611
null
patchbench-go-real-0006
go
real
swe-bench-live-multilang
ossf__scorecard-4728
ossf/scorecard
4feedb857ab8d82158aa9774bf8054df41992180
"Signed-Releases check ignores .sigstore.json signatures produced by sigstore/gh-action-sigstore-pyt(...TRUNCATED)
"diff --git a/docs/checks.md b/docs/checks.md\nindex 6d77d5df8c0..1885542805b 100644\n--- a/docs/che(...TRUNCATED)
"diff --git a/checks/signed_releases_test.go b/checks/signed_releases_test.go\nindex 45691358626..80(...TRUNCATED)
["github.com/ossf/scorecard/v5/checks/TestSignedRelease/Releases_with_assets_with_signed_artifacts-s(...TRUNCATED)
["github.com/ossf/scorecard/v5/checks/evaluation/TestLicense/Has_license_file_but_not_OSI/FSF_approv(...TRUNCATED)
[ "go test -v ./... -json > reports/go-test-results.json" ]
"def parser(log: str) -> dict[str, str]:\n import json\n import re\n\n results: dict[str, s(...TRUNCATED)
starryzhang/sweb.eval.x86_64.ossf_1776_scorecard-4728
null
patchbench-go-real-0070
go
real
swe-bench-live-multilang
twpayne__chezmoi-5016
twpayne/chezmoi
5ffd82cb395bc0c33ad885234a72bef108f8243c
"Allow to invoke `stdinIsATTY` in regular templates\n<!--\n\nchezmoi has a strict zero-tolerance pol(...TRUNCATED)
"> I can check that within the script itself, like `test -t 0` in Bash.\n\nWhat is the downside of u(...TRUNCATED)
"diff --git a/assets/chezmoi.io/docs/reference/templates/init-functions/stdinIsATTY.md b/assets/chez(...TRUNCATED)
"diff --git a/internal/cmd/testdata/scripts/templatefuncs.txtar b/internal/cmd/testdata/scripts/temp(...TRUNCATED)
["chezmoi.io/chezmoi/internal/cmd/TestScript","chezmoi.io/chezmoi/internal/cmd/TestScript/templatefu(...TRUNCATED)
["chezmoi.io/chezmoi/internal/chezmoi/TestTargetStateEntryApply/target_symlink_actual_file/0","chezm(...TRUNCATED)
[ "mkdir -p reports && go test -v -json ./... > reports/go-test-results.json" ]
"def parser(log: str) -> dict[str, str]:\n import json\n import re\n import xml.etree.Eleme(...TRUNCATED)
starryzhang/sweb.eval.x86_64.twpayne_1776_chezmoi-5016
null
patchbench-go-real-0062
go
real
swe-bench-live-multilang
open-telemetry__opentelemetry-go-6962
open-telemetry/opentelemetry-go
73b2cd39b0fb6ca97ee3fc2c495d79e0b72e2b11
"semconv: Add ErrorType function\n### Problem Statement\n\nThere is no easy way to set [error.type]((...TRUNCATED)
"Hi. I would like to work on this. Please, assign me if it is feasible\n@Bakhtiyar-Garashov, sure.\n(...TRUNCATED)
"diff --git a/CHANGELOG.md b/CHANGELOG.md\nindex 9be9fa06d99..7a1e084d5b9 100644\n--- a/CHANGELOG.md(...TRUNCATED)
"diff --git a/internal/tools/semconvkit/templates/error_type_test.go.tmpl b/internal/tools/semconvki(...TRUNCATED)
["go.opentelemetry.io/otel/semconv/v1.34.0::TestErrorType","go.opentelemetry.io/otel/semconv/v1.34.0(...TRUNCATED)
["go.opentelemetry.io/otel/baggage::TestBaggageParse/single_member_with_value_containing_equal_signs(...TRUNCATED)
[ "go test -json -v ./... > reports/go-test-results.json" ]
"def parser(log: str) -> dict[str, str]:\n # Improved parser for Go test -json output and textual(...TRUNCATED)
starryzhang/sweb.eval.x86_64.open-telemetry_1776_opentelemetry-go-6962
null
patchbench-go-real-0057
go
real
swe-bench-live-multilang
rook__rook-16336
rook/rook
540b0c8637e93f80e1a91d7ce94a87ba384179ba
"Possible CSI priorKeyCount status calculation issue\n<!-- **Are you in the right place?**\n1. For i(...TRUNCATED)
"@parth-gr would you have any spare time to look at this before v1.18 release?\n/assign\nThanks for (...TRUNCATED)
"diff --git a/pkg/operator/ceph/csi/secrets.go b/pkg/operator/ceph/csi/secrets.go\nindex 6bec80fc1b7(...TRUNCATED)
"diff --git a/pkg/operator/ceph/csi/secrets_test.go b/pkg/operator/ceph/csi/secrets_test.go\nindex d(...TRUNCATED)
["github.com/rook/rook/pkg/operator/ceph/csi::TestUpdateCsiClusterConfig/add_a_simple_mons_list","gi(...TRUNCATED)
["github.com/rook/rook/pkg/daemon/multus::Test_perNodeTypeCount_Equal/shared:1,other:2_vs_shared:2,o(...TRUNCATED)
[ "cd /testbed && go test ./... -v 2>&1 | tee test-output.log" ]
"def parser(log: str) -> dict[str, str]:\n # Robust Go test output parser.\n # Key features:\n(...TRUNCATED)
starryzhang/sweb.eval.x86_64.rook_1776_rook-16336
null
patchbench-go-real-0035
go
real
swe-bench-live-multilang
open-telemetry__opentelemetry-go-6393
open-telemetry/opentelemetry-go
c5e68b201067cb4e6b98a3e619927abc4a235e3e
"Trace SDK observability - batch processor metrics\nCommon requirement as this is an experimental fe(...TRUNCATED)
"diff --git a/CHANGELOG.md b/CHANGELOG.md\nindex c99e783f1e1..0cd7435aa45 100644\n--- a/CHANGELOG.md(...TRUNCATED)
"diff --git a/sdk/trace/batch_span_processor_test.go b/sdk/trace/batch_span_processor_test.go\nindex(...TRUNCATED)
["TestObservableExample/Delta","TestSpanLimit/AttributeCountLimit/ValidValue","TestNewBatchSpanProce(...TRUNCATED)
["Test_otTagsToOTelAttributesKindAndError/server","TestConvHeaders","TestUnquote/Half-open_starting_(...TRUNCATED)
[ "make test-verbose 2>&1 | tee test-output.log" ]
"def parser(log: str) -> dict[str, str]:\n import re\n results: dict[str, str] = {}\n # Mat(...TRUNCATED)
starryzhang/sweb.eval.x86_64.open-telemetry_1776_opentelemetry-go-6393
null
patchbench-go-real-0026
go
real
swe-bench-live-multilang
istio__istio-56962
istio/istio
2f99077494002e6b61e6eefa169ac12ae5bbfde0
"Gateway deployer runs deployment loop on each Gateway status update\nGateway status updates frequen(...TRUNCATED)
"@howardjohn are you thinking short circuiting the Reconcile in [deploymentcontroller.go](https://gi(...TRUNCATED)
"diff --git a/pilot/pkg/config/kube/gateway/deploymentcontroller.go b/pilot/pkg/config/kube/gateway/(...TRUNCATED)
"diff --git a/pilot/pkg/config/kube/gateway/deploymentcontroller_test.go b/pilot/pkg/config/kube/gat(...TRUNCATED)
["istio.io/istio/pilot/pkg/config/kube/gateway::TestConvertResources/delegated","istio.io/istio/pilo(...TRUNCATED)
["istio.io/istio/pilot/pkg/model/kstatus::TestUpdateConditionIfChanged/unchanged","istio.io/istio/pi(...TRUNCATED)
[ "/usr/local/go/bin/go test -v ./... -json > reports/go-test-results.json" ]
"def parser(log: str) -> dict[str, str]:\n import re\n import json\n\n results: dict[str, s(...TRUNCATED)
starryzhang/sweb.eval.x86_64.istio_1776_istio-56962
null
patchbench-go-real-0022
go
real
swe-bench-live-multilang
kubernetes-sigs__controller-runtime-3247
kubernetes-sigs/controller-runtime
252af6420feba970d0020ba4c029a0a1ce3f9688
"Context cancellation causes a goroutine leak (possibly intermittent)\n### Problem statement\n\nWhen(...TRUNCATED)
"Hi @pmalek, are you still seeing this issue? Which controller-runtime version are you using? I did (...TRUNCATED)
"diff --git a/pkg/manager/manager.go b/pkg/manager/manager.go\nindex c3ae317b04..09207fa302 100644\n(...TRUNCATED)
"diff --git a/pkg/manager/manager_test.go b/pkg/manager/manager_test.go\nindex 247a33f9dc..25bc02c3e(...TRUNCATED)
[ "sigs.k8s.io/controller-runtime/pkg/manager::TestSource" ]
["sigs.k8s.io/controller-runtime/pkg/cache::TestDefaultOpts/DefaultNamespace_label_selector_gets_def(...TRUNCATED)
[ "go test -json -v ./... > reports/go-test-results.json" ]
"def parser(log: str) -> dict[str, str]:\n import json\n results: dict[str, str] = {}\n for(...TRUNCATED)
starryzhang/sweb.eval.x86_64.kubernetes-sigs_1776_controller-runtime-3247
null
patchbench-go-real-0108
go
real
swe-bench-live-multilang
go-vikunja__vikunja-2574
go-vikunja/vikunja
457ffbfe5133e2b42343c542092f027b8103b681
"When finishing a recurring task on the kanban board the task should move to the to-do column\n### P(...TRUNCATED)
"diff --git a/pkg/models/kanban_task_bucket.go b/pkg/models/kanban_task_bucket.go\nindex 35c623a0f4.(...TRUNCATED)
"diff --git a/pkg/models/kanban_task_bucket_test.go b/pkg/models/kanban_task_bucket_test.go\nindex a(...TRUNCATED)
["code.vikunja.io/api/pkg/models/TestTask_Update","code.vikunja.io/api/pkg/models/TestTask_Update/re(...TRUNCATED)
["code.vikunja.io/api/pkg/webtests/TestTaskComments/Create/Permissions_check","code.vikunja.io/api/p(...TRUNCATED)
[ "mkdir -p reports && go test -json ./... > reports/go-test-results.json" ]
"def parser(log: str) -> dict[str, str]:\n import json\n import re\n\n results: dict[str, s(...TRUNCATED)
starryzhang/sweb.eval.x86_64.go-vikunja_1776_vikunja-2574
null
patchbench-go-real-0008
go
real
swe-bench-live-multilang
graphql-go__graphql-727
graphql-go/graphql
f2a7e55af5fff16f47b3c807b9c8d318526f7ea5
"`{introspection}`: Grammar fixes.\n#### Actual\nCurrently `introspection` pkg contains grammar miss(...TRUNCATED)
"If you would like me to solve this issue, either tag me in a comment or check this box:\n- [ ] Solv(...TRUNCATED)
"diff --git a/introspection.go b/introspection.go\nindex 51feb42d..31e802d7 100644\n--- a/introspect(...TRUNCATED)
"diff --git a/introspection_test.go b/introspection_test.go\nindex c0e62bf1..735144c2 100644\n--- a/(...TRUNCATED)
["github.com/graphql-go/graphql::TestIntrospection_ExposesDescriptionsOnTypesAndFields","github.com/(...TRUNCATED)
["github.com/graphql-go/graphql::TestIsTypeSubTypeOf_SameReferenceIsSubtype","github.com/graphql-go/(...TRUNCATED)
[ "go test -v ./... -json > test-output.json" ]
"def parser(log: str) -> dict[str, str]:\n import json\n import re\n\n results: dict[str, s(...TRUNCATED)
starryzhang/sweb.eval.x86_64.graphql-go_1776_graphql-727
null
End of preview.

YAML Metadata Warning:empty or missing yaml metadata in repo card

Check out the documentation for more information.

PatchBench

A multi-language benchmark for evaluating whether an LLM can fix a bug correctly and without introducing new bugs. Each row presents a real issue report plus a large code context (~500–1500 lines); the model under test produces a patch; a per-row test suite then checks that (a) previously failing tests now pass and (b) the rest of the suite stays green.

Status: v0 spec. Rows are being built.

Row schema

Field Type In tasks config Description
id string ✅ patchbench-{lang}-{source}-{nnnn}
language string ✅ python | javascript | typescript | go | java | rust
source string ✅ real (mined bug-fix PR) | synthetic (injected bug)
origin string ✅ Source dataset, e.g. swe-bench-live-multilang, swesmith-go
repo string ✅ GitHub repo the code came from
base_commit string ✅ Commit the buggy state is taken at
issue_text string ✅ Issue/PR report the model must diagnose from
context_files list of {path, content} ✅ ~500–1500 lines total incl. the buggy file(s), callers, and related modules
buggy_files list of string ✅ Paths the fix is expected to touch
test_files list of {path, content} ✅ Test patch materialized as files
test_cmds list of string ✅ Commands that run the tests for this row (language-native)
env_setup_cmds list of string ❌ (harness only) Install/build commands needed before test_cmds
fail_to_pass list of string ❌ Tests that MUST fail pre-fix and pass post-fix
pass_to_pass list of string ❌ Tests that MUST stay passing post-fix (the regression check)
gold_patch string ❌ Reference fix, for harness/self-eval only

Two configs prevent leakage:

  • tasks (public eval surface): everything an agent sees — issue, context, tests.
  • gold (harness-only): fail_to_pass, pass_to_pass, gold_patch, env_setup_cmds.

Verification contract (every released row)

  1. Pre-fix: apply test_files to the repo at base_commit → every fail_to_pass test fails; every pass_to_pass test passes.
  2. Post-fix: apply gold_patch → every fail_to_pass and pass_to_pass test passes.
  3. Determinism: each check is run twice; any row with a flaky test is dropped.
  4. No-leak check: issue text must not name the exact function/line to change (auto-screen + spot manual review).

A model's fix is scored: resolved iff all fail_to_pass pass AND all pass_to_pass still pass. Rows also record per-test results so partial behavior (fix works but broke something) is visible — that is the "introduced a new bug" signal.

Sourcing plan (target ~1,500–2,000 rows)

Language Real source Synthetic top-up Target
Python SWE-Gym + SWE-bench Verified/train SWE-smith-py ~550
Rust SWE-bench-Live rust + rustbench SWE-smith-rs ~300
Java SWE-bench-Live java + Multi-SWE-bench java SWE-smith-java ~300
Go SWE-bench-Live go + Multi-SWE-bench go SWE-smith-go ~250
JS/TS SWE-bench-Live js+ts + Multi-SWE-bench SWE-smith-js/ts ~400

Context expansion is PatchBench's own contribution: no upstream dataset ships sliced 500–1500-line windows. We materialize the repo at base_commit, compute the failing-test dependency closure, and slice the modules inside it, filtering out rows whose closure exceeds the window.

Harness

Language-native runner (no Docker required, adapted from the SWE-bench grading loop + microsoft/RepoLaunch per-language build/test recipes): clone at base_commit → env_setup_cmds → apply test_files (+ candidate patch) → run test_cmds → parse per-test results → grade against F2P/P2P. Per-repo env caching makes verification tractable.

License audit

Upstream dataset licenses: SWE-bench-Live (MIT), SWE-Gym (MIT), SWE-smith (MIT). Multi-SWE-bench / PrimeIntellect rows are CC0/“other” (ByteDance community terms) — those rows are re-verified from original permissively-licensed repos before inclusion, or excluded.
Downloads last month
12
Total file size:
83 MB