We submitted a ~100-page full report on our assessment to Anthropic.
For the public version, Anthropic redacted a majority of the evidence we found most informative. As a consequence, the public must rely on our judgment more than we consider ideal.
0
0 comments