We have 3 of 15 papers
We haven’t gathered this author’s papers yet. Follow them and we’ll fetch their work.
Not the right person? Other researchers publish under this name.
Results show that inoculation with a learned association introduced via midtraining can shape selective generalisation, and more work is needed before this approach can become a load-bearing component in a developer's safety framework.
Evaluating 13-21 models across six presentation operationalizations and four task-domain operationalizations suggests that, despite confounds, some models possess practical SGTR capabilities, and that SGTR should be monitored and considered in the design of safety-critical AI applications.
Post-training alignment is often shallow, eroding under fine-tuning. It remains untested as to whether constitutional midtraining interventions can produce durable alignment when cleanly isolated from post-training. We build a 394M-token constitutional corpus from Anthropic's Constitution and apply constitutional midtr...
We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.
