Hugging Face's logo

Datasets:
abhid1234
/
agentroute-fixtures

Tags:
ai-routing
llm-observability
decision-receipts
Synthetic
fixtures
License:
Dataset card Files
xet
 Community
Dataset Viewer
Duplicate
The dataset viewer is not available for this split.
Cannot load the dataset split (in streaming mode) to extract the first rows.
Error code:   StreamingRowsError
Exception:    CastError
Message:      Couldn't cast
route_version: string
record_type: string
route_id: string
created_at: string
task: struct<type: string, fingerprint: string>
  child 0, type: string
  child 1, fingerprint: string
router: struct<name: string, policy_id: string>
  child 0, name: string
  child 1, policy_id: string
source: struct<kind: string, fidelity: string, event_id: string>
  child 0, kind: string
  child 1, fidelity: string
  child 2, event_id: string
candidates: list<item: struct<id: string, model: string, provider: string, eligible: bool, estimates: struct<qua (... 133 chars omitted)
  child 0, item: struct<id: string, model: string, provider: string, eligible: bool, estimates: struct<quality: doubl (... 121 chars omitted)
      child 0, id: string
      child 1, model: string
      child 2, provider: string
      child 3, eligible: bool
      child 4, estimates: struct<quality: double, latency_ms: int64, cost_usd: double>
          child 0, quality: double
          child 1, latency_ms: int64
          child 2, cost_usd: double
      child 5, scores: struct<overall: double, quality: double, latency: double, cost: double>
          child 0, overall: double
          child 1, quality: double
          child 2, latency: double
          child 3, cost: double
criteria: struct<min_quality: double>
  child 0, min_quality: double
selection: struct<candidate_id: string, reason: string>
  child 0, candidate_id: string
  child 1, reason: string
extensions: struct<demo_fixture: string, openrouter: struct<requested: string, strategy: string, attempt: int64> (... 1 chars omitted)
  child 0, demo_fixture: string
  child 1, openrouter: struct<requested: string, strategy: string, attempt: int64>
      child 0, requested: string
      child 1, strategy: string
      child 2, attempt: int64
observation_id: string
observed_at: string
outcome: struct<status: string, actual_model: string, actual_provider: string, latency_ms: int64, cost_usd: d (... 72 chars omitted)
  child 0, status: string
  child 1, actual_model: string
  child 2, actual_provider: string
  child 3, latency_ms: int64
  child 4, cost_usd: double
  child 5, quality: double
  child 6, metadata: struct<evaluator: struct<id: string>>
      child 0, evaluator: struct<id: string>
          child 0, id: string
version: string
id: string
weights: struct<quality: double, latency: double, cost: double>
  child 0, quality: double
  child 1, latency: double
  child 2, cost: double
to
{'id': Value('string'), 'version': Value('string'), 'criteria': {'max_cost_usd': Value('float64'), 'max_latency_ms': Value('int64'), 'min_quality': Value('float64')}, 'weights': {'quality': Value('float64'), 'latency': Value('float64'), 'cost': Value('float64')}}
because column names don't match
Traceback:    Traceback (most recent call last):
                File "/src/services/worker/src/worker/utils.py", line 147, in get_rows_or_raise
                  return get_rows(
                      dataset=dataset,
                  ...<4 lines>...
                      column_names=column_names,
                  )
                File "/src/libs/libcommon/src/libcommon/utils.py", line 272, in decorator
                  return func(*args, **kwargs)
                File "/src/services/worker/src/worker/utils.py", line 127, in get_rows
                  rows_plus_one = list(itertools.islice(safe_iter(ds, dataset=dataset), rows_max_number + 1))
                File "/src/services/worker/src/worker/utils.py", line 483, in safe_iter
                  yield from ds.decode(False) if ds.features else ds
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2840, in __iter__
                  for key, example in ex_iterable:
                                      ^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2373, in __iter__
                  for key, pa_table in self._iter_arrow():
                                       ~~~~~~~~~~~~~~~~^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2398, in _iter_arrow
                  for key, pa_table in self.ex_iterable._iter_arrow():
                                       ~~~~~~~~~~~~~~~~~~~~~~~~~~~~^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 536, in _iter_arrow
                  for key, pa_table in iterator:
                                       ^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 419, in _iter_arrow
                  for key, pa_table in self.generate_tables_fn(**gen_kwags):
                                       ~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^
                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
              route_version: string
              record_type: string
              route_id: string
              created_at: string
              task: struct<type: string, fingerprint: string>
                child 0, type: string
                child 1, fingerprint: string
              router: struct<name: string, policy_id: string>
                child 0, name: string
                child 1, policy_id: string
              source: struct<kind: string, fidelity: string, event_id: string>
                child 0, kind: string
                child 1, fidelity: string
                child 2, event_id: string
              candidates: list<item: struct<id: string, model: string, provider: string, eligible: bool, estimates: struct<qua (... 133 chars omitted)
                child 0, item: struct<id: string, model: string, provider: string, eligible: bool, estimates: struct<quality: doubl (... 121 chars omitted)
                    child 0, id: string
                    child 1, model: string
                    child 2, provider: string
                    child 3, eligible: bool
                    child 4, estimates: struct<quality: double, latency_ms: int64, cost_usd: double>
                        child 0, quality: double
                        child 1, latency_ms: int64
                        child 2, cost_usd: double
                    child 5, scores: struct<overall: double, quality: double, latency: double, cost: double>
                        child 0, overall: double
                        child 1, quality: double
                        child 2, latency: double
                        child 3, cost: double
              criteria: struct<min_quality: double>
                child 0, min_quality: double
              selection: struct<candidate_id: string, reason: string>
                child 0, candidate_id: string
                child 1, reason: string
              extensions: struct<demo_fixture: string, openrouter: struct<requested: string, strategy: string, attempt: int64> (... 1 chars omitted)
                child 0, demo_fixture: string
                child 1, openrouter: struct<requested: string, strategy: string, attempt: int64>
                    child 0, requested: string
                    child 1, strategy: string
                    child 2, attempt: int64
              observation_id: string
              observed_at: string
              outcome: struct<status: string, actual_model: string, actual_provider: string, latency_ms: int64, cost_usd: d (... 72 chars omitted)
                child 0, status: string
                child 1, actual_model: string
                child 2, actual_provider: string
                child 3, latency_ms: int64
                child 4, cost_usd: double
                child 5, quality: double
                child 6, metadata: struct<evaluator: struct<id: string>>
                    child 0, evaluator: struct<id: string>
                        child 0, id: string
              version: string
              id: string
              weights: struct<quality: double, latency: double, cost: double>
                child 0, quality: double
                child 1, latency: double
                child 2, cost: double
              to
              {'id': Value('string'), 'version': Value('string'), 'criteria': {'max_cost_usd': Value('float64'), 'max_latency_ms': Value('int64'), 'min_quality': Value('float64')}, 'weights': {'quality': Value('float64'), 'latency': Value('float64'), 'cost': Value('float64')}}
              because column names don't match

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.

AgentRoute Fixtures

Reference fixtures for AgentRoute — a vendor-neutral evidence and governance layer for AI routing decisions. AgentRoute never proxies model traffic, never picks a model, and never applies a policy on its own; it turns routing decisions into typed, replayable receipts.

Illustrative, synthetic data. Every file here is authored or deterministically generated offline from bundled fixtures. Nothing is a live model benchmark or provider-performance claim, and no real traffic, prompts, or completions are included. Model and provider names in the examples (e.g. frontier-review-model, provider-a) are illustrative.

What's inside

Directory Contents
examples/ Route ledgers (*.route.jsonl — the decision/observation trajectories), preregistered experiment protocols, policies, quality gates, drift/SLO/scenario configs, and saved-log import fixtures for OpenRouter, LiteLLM, Portkey, Vercel AI Gateway, Cloudflare AI Gateway, and Braintrust
route-conformance/ The conformance corpus: valid and deliberately invalid receipt cases used by npm run conformance (5 cases)
schema/ Draft 2020-12 JSON Schema for decision and observation receipts
proof-pack/ A complete generated proof pack: 31 linked artifacts (replay receipts, preregistered experiment decision, quality gate, five-target dry-run promotion dossier, drift/SLO/outage review, hash-chained reliability timeline) verified to a single root hash
manifest.json Path, size, and SHA-256 for every file in this dataset

The receipt rail

Every route is organized as Requested → Selected → Observed → Proposed — and "Proposed" is deliberately a labeled prediction over recorded routing-time scores, never called "better" until executed and measured.

Reproduce everything locally

No accounts, no keys, no network:

git clone https://github.com/abhid1234/AgentRoute && cd AgentRoute
npm ci --ignore-scripts && npm run build
node dist/cli.js proof run --out local/proof-pack
node dist/cli.js proof verify local/proof-pack

Links

Downloads last month
7
Total file size:
542 kB