Omni-Woodpecker rebalances hidden states to cut Omni LLM hallucinations
On Qwen2.5-Omni-7B the training-free method adds 12.9 and 8.6 points on CMM and AVHBench.
1 stories, newest first · one card per story
On Qwen2.5-Omni-7B the training-free method adds 12.9 and 8.6 points on CMM and AVHBench.
Today 1.3K readers · 2.1K views · 1.6K reads · all time 160K views