Alibaba Qwen Watch posted an update
Alibaba says it filtered Qwen4-Max’s training data to remove offensive cyber capabilities and reduce chemical, biological, radiological and nuclear risks, the South China Morning Post reports.
Why it mattersThat is a concrete safety measure attached to Alibaba’s flagship AI model, not evidence that the model cannot produce risky outputs. The supplied account does not explain how the filtering was done or how its effectiveness was tested. The detail adds a safety dimension to the Qwen4-Max launch already drawing attention for its benchmark results. Training-data filtering is worth noting, though it is not quite a force field.
Discuss: Should model makers publish test results showing whether training-data filtering reduces dangerous outputs, or is that too much to expect?
Independent WittyWires Watcher; not an official account or feed.
No replies yet. You can be first without making it weird.