Evaluation report — replies

Customer
bookshop
Suite
replies
This run
dev — 2026-10-28T12:55:30.940004+00:00
Reference
dev — 2026-10-28T12:25:33.365220+00:00
Code version
9bbc16222984c8c46b4da68722c10b15c0f5f1d0
Reference approved
2026-10-28T12:28:00.539388+00:00
Reference key
2026-10-28T12-25-33-365220-00-00-efe01136327fd179

Did it get worse? No

Nothing got worse compared with the reference. 2 checks are on the line: the bands they measured cover their thresholds. Every case could be judged. No case is suspended. The suite is unchanged from the reference. The files under test are the same as the reference. The system under test answered under the same configuration as the reference.

What was under test

The files under test are the same as the reference.

How it was set up

Configured as in the reference.

ParameterThis runReference
max_tokens200200
modelclaude-haiku-4-5claude-haiku-4-5
provideranthropicanthropic
resolved_modelclaude-haiku-4-5-20251001claude-haiku-4-5-20251001

How the judge was set up

Configured as in the reference.

ParameterThis runReference
judgeanthropic/claude-haiku-4-5anthropic/claude-haiku-4-5
max_tokens400400
modelclaude-haiku-4-5claude-haiku-4-5
provideranthropicanthropic
resolved_modelclaude-haiku-4-5-20251001claude-haiku-4-5-20251001
What got worse (0)

Nothing in this section.

What could not be judged (0)

Nothing in this section.

What is set aside (0)

Nothing in this section.

What was added or removed (0)

Nothing in this section.

What got better (0)

Nothing in this section.

What stayed the same (10)
CaseCheckWhat happenedWhy
gift-wraplevenshteinScore unchanged at 0.191404.mean of 5 samples (0.174107, 0.190722, 0.210227, 0.198795, 0.183168)
gift-wrapllm_rubricScore unchanged at 0.540000.mean of 5 samples (0.616667, 0.666667, 0.433333, 0.450000, 0.533333)
opening-hourslevenshteinScore unchanged at 0.196253.mean of 5 samples (0.212389, 0.205479, 0.200893, 0.182927, 0.179577)
opening-hoursllm_rubricScore unchanged at 0.343333.mean of 5 samples (0.350000, 0.333333, 0.300000, 0.433333, 0.300000)
order-a-booklevenshteinScore unchanged at 0.271600.mean of 5 samples (0.228137, 0.270270, 0.299517, 0.277778, 0.282297)
order-a-bookllm_rubricScore unchanged at 0.540000.mean of 5 samples (0.683333, 0.816667, 0.533333, 0.366667, 0.300000)
price-checklevenshteinScore unchanged at 0.194523.mean of 5 samples (0.189516, 0.216590, 0.179211, 0.182573, 0.204724)
price-checkllm_rubricScore unchanged at 0.310000.mean of 5 samples (0.300000, 0.333333, 0.250000, 0.333333, 0.333333)
signed-copieslevenshteinScore unchanged at 0.197552.mean of 5 samples (0.206349, 0.206226, 0.190299, 0.193916, 0.190972)
signed-copiesllm_rubricScore unchanged at 0.443333.mean of 5 samples (0.300000, 0.716667, 0.466667, 0.300000, 0.433333)
Where the judge placed the calibration answers (1)
CaseCheckResultDeclared bandWhy
calibration-half-rightllm_rubric0.743333 inside0.100000–0.900000mean of 5 samples (0.733333, 0.733333, 0.766667, 0.716667, 0.766667)