Merge pull request rllm-org#162 from UKGovernmentBEIS/80_swe_bench_test_fix

MattFisher · web-flow · commit f18e8704322c · 2025-01-20T10:02:44.000+11:00
Use a working SWE Bench sample for the unit test
diff --git a/src/inspect_evals/swe_bench/tests/swe_bench_tests.py b/src/inspect_evals/swe_bench/tests/swe_bench_tests.py
@@ -20,7 +20,7 @@
 
 
 SWE_BENCH_SPLIT = ("princeton-nlp/SWE-bench_Verified", "test")
-TEST_INSTANCE_ID = "django__django-10097"
+TEST_INSTANCE_ID = "django__django-17087"
 
 
 @solver
