Eval receipt
Demo data
GPT OSS 120B on 3 math questions.
Scored 1 of 3 answers.
Run 1 of 3- Registered before the run
- The model, questions and grader.
- Model
- GPT OSS 120B
- Test set
- 3 math questions
- Grader
- Compares the final number with the expected answer.
- Cost
- $0.00 claimed for this run
Why did it change?
The runner tested GPT OSS 120B on 3 math questions and reported 1 correct answers.
Who approved it?
No approval record is attached.
What did it cost?
$0.00 claimed for this run. No bill backs it up yet.
How do we know?
Bitcoin shows when this saved result already existed. The score is the runner's claim.
Entries
5 entries on 1 day-
October 7
On October 7, the runners recorded 5 test runs. Earlier scores remain visible.
4:24 PM UTCA runner reported 0 of 3 answers correct for Qwen 3.6 27B in run 1 of 2.
The score and cost are reported by the runner.
Open Qwen 3.6 27B run 1 of 2$0.00 Claimed4:25 PM UTCA runner reported 1 of 3 answers correct for GPT OSS 120B in run 1 of 3.
The score and cost are reported by the runner.
$0.00 Claimed4:25 PM UTCA runner reported 1 of 3 answers correct for GPT OSS 120B in run 2 of 3.
The score and cost are reported by the runner.
Open GPT OSS 120B run 2 of 3$0.00 Claimed8:33 PM UTCA runner reported 3 of 3 answers correct for GPT OSS 120B in run 3 of 3.
The score and cost are reported by the runner.
Open GPT OSS 120B run 3 of 3$0.00 Claimed8:35 PM UTCA runner reported 2 of 3 answers correct for Qwen 3.6 27B in run 2 of 2.
The score and cost are reported by the runner.
Open Qwen 3.6 27B run 2 of 2$0.00 Claimed
Proof
Can't be changed without it showing.
Stamped on Bitcoin at 4:53 PM UTC, October 7.
That stamp covers this saved test result.