VPSG corrects predictable directional coordinate biases in MLLMs by shuffling visual positional encodings to isolate unconditioned tendencies and steering digit decoding with a lightweight finite-state machine, yielding accuracy gains on ScreenSpot-Pro without retraining.
Insights into llm long- context failures: when transformers know but don’t tell
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2025 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Mitigating Coordinate Prediction Bias from Positional Encoding Failures
VPSG corrects predictable directional coordinate biases in MLLMs by shuffling visual positional encodings to isolate unconditioned tendencies and steering digit decoding with a lightweight finite-state machine, yielding accuracy gains on ScreenSpot-Pro without retraining.