frontier models
One Forward Pass Had to Answer—and Know When Not To
YOPO reconstructed the pre-steering residual stream so a frozen model could improve reasoning without blinding its own abstention signal.
Summary
YOPO reconstructed the pre-steering residual stream so a frozen model could improve reasoning without blinding its own abstention signal.
Writing a steering intervention into the residual stream changed the very signal used to decide whether evidence was sufficient. YOPO trained a small reconstructor on paired clean and steered states, then read the fixed sufficiency direction from that reconstruction. On reported Qwen2.5 backbones it combined answering, steering and abstention in one pass and outperformed the two-pass reference, though the authors also found and disclosed a surface artifact in one benchmark construction.
Why it matters
YOPO reconstructed the pre-steering residual stream so a frozen model could improve reasoning without blinding its own abstention signal.
Limits and context
No additional limitation was separately recorded.
Key claims
YOPO reconstructed the pre-steering residual stream so a frozen model could improve reasoning without blinding its own abstention signal.
Evidence: source-2026-08-17-006
Sources
- arXiv preprint 2608.14465arXiv · primary research
Corrections
No corrections have been recorded for this story.