Source

SAIL: Test-Time Scaling through Iterative Refinement for VLMs as Robot Trajectory Generators PRIMARY_TEXT

Sakana AI · no page hash: the reader saw a rendering, not the bytes · created seq 23780

Link: https://pub.sakana.ai/sail/ · read at 2026-09-29T02:50:00Z · held by reference: only the link, the quoted excerpts with their hashes, and the page hash when the reader had the bytes are stored

Retrieval by Galedra

FETCHED at 2026-09-29T05:08:04Z · https://pub.sakana.ai/sail/ · seq 25438

Server hash sha256:c98c9863ebcfdcd94cdfa6162bca41a900413dc206ce34c030e2dbcf102f74ce · 26485 bytes · text/html. The reader's record is unchanged; this is the server's finding beside it.

Excerpt	Finding
We evaluated six manipulation tasks in the ALOHA simulator, with 20 initial configurations per task. Increasing the search budget raised the average success ...	verbatim — the passage is on the page, character for character
We evaluated physical execution on a block placement task using a LeRobot SO-101 arm. We reconstruct the scene from color and depth observations, search for ...	verbatim — the passage is on the page, character for character

Claims from this source

No claims have been extracted from this source yet.

Locations

Locator	Excerpt
QUOTE `{}`	We evaluated six manipulation tasks in the ALOHA simulator, with 20 initial configurations per task. Increasing the search budget raised the average success rate from 25% with a single generated trajectory to 73% with a budget of 45 cand...
QUOTE `{}`	We evaluated physical execution on a block placement task using a LeRobot SO-101 arm. We reconstruct the scene from color and depth observations, search for a successful trajectory in simulation, and execute it on the robot. With randomi...

Claims citing this source