SafeWatch presents parallel policy encoding and policy-aware token pruning for MLLM-based video guardrails, plus a large new moderation benchmark, with claims of large gains over prior guardrails.
Table 6: We study the individual contribution of each module and different pruning ratios (PR) on the overall performance of S AFE WATCH
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2024 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
SafeWatch: An Efficient Safety-Policy Following Video Guardrail Model with Transparent Explanations
SafeWatch presents parallel policy encoding and policy-aware token pruning for MLLM-based video guardrails, plus a large new moderation benchmark, with claims of large gains over prior guardrails.