A radar-only 69-token interface to frozen LLMs achieves 98.13% proposal recall and shows, under matched controls, that aligned language supervision provides no stable direct-head benefit.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Radar4D-VLM: Proposal-Grounded Temporal 4D Radar Reasoning Across Frozen Language Models
A radar-only 69-token interface to frozen LLMs achieves 98.13% proposal recall and shows, under matched controls, that aligned language supervision provides no stable direct-head benefit.