max_tokens_up
Generation budget raised from 256 to 1024 tokens
Guard successfull suite successchange categories: generation_config4 checks run, impact analysis 0.3 s
Guard, live
Live runs are turned off on this public site. The results below are recorded; live runs happen on the owner's runner.
Checks: targeted selection vs full suite
Recorded in RQ3 (2026-10-06). Its perf-gate figures predate the gate's warmup fix; see the corrections; the re-measured numbers are in the repair timeline below.
| check | selected by Guard | result | s | detail |
|---|---|---|---|---|
| unit:tests/test_retriever.py | no: exercises nothing this change touches | pass | 1.6 | 2 passed in 0.52s |
| unit:tests/test_prompts.py | no: exercises nothing this change touches | pass | 1.6 | 2 passed in 0.58s |
| check:citation_required | yes | pass | 10.9 | citation contract met in 1/34 answers on the PR vs 0/34 on base (+2.9 pts; budget -5 pts; p=1.00); via full answer path (incl. post-processing) |
| judge:ungrounded_answer | yes | skip | 0.0 | no judge model key configured (BYOK); judge not run |
| check:prompt_prefix_stable | no: exercises nothing this change touches | pass | 3.8 | requests share 3954 leading characters (base 3954, need 3558) |
| check:context_budget | no: exercises nothing this change touches | pass | 7.7 | largest prompt 980 tokens (base 980) against a budget of 3840; median 964 vs 964 |
| security:approval_gate | yes | pass | 1.4 | 1 approval test(s) (generated from the base revision): 1 passed in 0.30s |
| bench:chat_perf_gate | yes | pass | 44.3 | median p95 TTFT 238.4 -> 226.4 ms (-5.0%, pass; 3 runs each, runs overlap); prefix-cache hit 98% -> 98% |
Check run as posted to GitHub (markdown)
# No regressions in 4 targeted checks conclusion: success **Ran 4 of 8 suite items** selected from the behavior-to-code graph. Change categories: generation_config. ### Quality | check | result | detail | |---|---|---| | `check:citation_required` | **PASS** | citation contract met in 1/34 answers on the PR vs 0/34 on base (+2.9 pts; budget -5 pts; p=1.00); via full answer path (incl. post-processing) | | `judge:ungrounded_answer` | **SKIP** | no judge model key configured (BYOK); judge not run | ### Performance | check | result | detail | |---|---|---| | `bench:chat_perf_gate` | **PASS** | median p95 TTFT 238.4 -> 226.4 ms (-5.0%, pass; 3 runs each, runs overlap); prefix-cache hit 98% -> 98% | ### Security | check | result | detail | |---|---|---| | `security:approval_gate` | **PASS** | 1 approval test(s) (generated from the base revision): 1 passed in 0.30s | <details><summary>Not run for this change</summary> - `unit:tests/test_retriever.py`: exercises nothing this change touches - `unit:tests/test_prompts.py`: exercises nothing this change touches - `check:prompt_prefix_stable`: exercises nothing this change touches - `check:context_budget`: exercises nothing this change touches </details> ### What changed in the graph - `config_key:config/rag.yaml#generation.max_tokens`: `value` 256 -> 1024