bench: unpack 3-tuple from get_benchmark_params

aca-model's `get_benchmark_params` returns three values (likely a
moments tuple alongside the two pre-existing entries); the benchmark
file was stuck on the older 2-tuple shape. aca-model's own
`test_benchmark.py` runs against the up-to-date in-tree helper, so the
drift only surfaces here once pylcm's benchmark CI pulls a fresh
aca-model.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

Author: hmgaudecker

benchmarks/bench_aca_baseline.py

@@ -71,7 +71,7 @@ def _build() -> tuple[object, object, object]:
         n_subjects=_N_SUBJECTS,
         pref_type_grid=DiscreteGrid(BenchmarkPrefType),
     )
-    _, model_params = get_benchmark_params(model=model)
+    _, _, model_params = get_benchmark_params(model=model)
     initial_conditions = get_benchmark_initial_conditions(
         model=model, n_subjects=_N_SUBJECTS, seed=0
     )