sensor-placement-shift-rare-activity

Task gallery / Machine learning / Version 1

Reference solution (oracle): Pass
No-op resistance check: Pass
QA verdict: Not recorded
Aggregate eligibility: Eligible version

The task

Train a detector that flags a rare activity in short motion windows and predict, for each held out window, the probability that it is that activity. Training windows come from a sensor on the arm, while graded windows come from a leg sensor on different people, so the model must transfer across a change of sensor placement rather than memorize one location. The rare activity is about a tenth of the windows and share…

Excerpt only; not a complete executable task specification. Task ID task_8505ce85fff14d50 · version 1. Selection status: not fully recorded.

Results by model and effort

This exact task version only. Every rate is passes / valid runs with a Wilson 95% interval; n≤2 remains counts only. Macro aggregation does not apply within a single task. Baselines are summarized in the quality gates, not ranked. A passed no-op resistance check means that doing nothing did not solve the task.

Current evidence

Model / effort	Verifier result [95% Wilson]	Diagnostic labels / valid	HARNESS_ERROR · excluded	Outcome classes
Claude Opus 5 · default (unverified)	3/4 · 75.0% [30.1%, 95.4%] · preliminary	4 / 4	3	GOOD_FAILURE: 1 BAD_FAILURE: 0 GOOD_SUCCESS: 3 BAD_SUCCESS: 0

Legacy evidence

Model / effort	Verifier result [95% Wilson]	Diagnostic labels / valid	HARNESS_ERROR · excluded	Outcome classes

Every recorded run

11 recorded trials, including controls. Diagnostic labels do not replace verifier reward. Labels and quality checks are source declarations, not independently attested here. Internal evidence remains private; a missing public URL is shown explicitly. Records with invalid or absent rewards are not scored.

Run / model	Date / source	Reward / eligibility	Diagnostic class / source	Resources	Evidence
`trial_6c54eb6b545e4f7d` oracle@@	Oct 11, 2026 console	1 · valid	No diagnostic label derived	Tokens: — Turns: — Cost: —	Not publicly available
`trial_a3665149b07c41a5` nop@@	Oct 11, 2026 console	0 · valid	No diagnostic label derived	Tokens: — Turns: — Cost: —	Not publicly available
`trial_2514544714c34e06` Claude Opus 5 · default (unverified)	Oct 11, 2026 console	No reward · excludedno_reward	HARNESS_ERROR derived	Tokens: — Turns: — Cost: —	Not publicly available
`trial_37936de555d04d14` Claude Opus 5 · default (unverified)	Oct 11, 2026 console	No reward · excludedno_reward	No diagnostic label No diagnostic source	Tokens: — Turns: — Cost: —	Not publicly available
`trial_510f0f240d034f3a` Claude Opus 5 · default (unverified)	Oct 11, 2026 console	No reward · excludedno_reward	HARNESS_ERROR derived	Tokens: — Turns: — Cost: —	Not publicly available
`trial_a12a034e25e24945` Claude Opus 5 · default (unverified)	Oct 11, 2026 console	No reward · excludedno_reward	HARNESS_ERROR derived	Tokens: — Turns: — Cost: —	Not publicly available
`trial_dfa4c5a80d004e35` Claude Opus 5 · default (unverified)	Oct 11, 2026 console	No reward · excludedno_reward	No diagnostic label No diagnostic source	Tokens: — Turns: — Cost: —	Not publicly available
`trial_74b9b456c09f4b1b` Claude Opus 5 · default (unverified)	Oct 11, 2026 console	0 · valid	GOOD_FAILURE frozen-classifier	Tokens: — Turns: — Cost: —	Not publicly available
`trial_3621acd81dd147a0` Claude Opus 5 · default (unverified)	Oct 11, 2026 console	1 · valid	GOOD_SUCCESS frozen-classifier	Tokens: — Turns: — Cost: —	Not publicly available
`trial_da0af232e9e2425e` Claude Opus 5 · default (unverified)	Oct 11, 2026 console	1 · valid	GOOD_SUCCESS frozen-classifier	Tokens: 43,475 Turns: 51 Cost: $3.2275	Not publicly available
`trial_9a1f87112c5549bc` Claude Opus 5 · default (unverified)	Oct 11, 2026 console	1 · valid	GOOD_SUCCESS frozen-classifier	Tokens: — Turns: — Cost: —	Not publicly available