Author

Muhammad Awais Bin Adil

1 paper indexed here

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.

#machine learning Preprint Jul 2026

No Usable Linear"Capitulation Direction"in Two Small LLMs: A Validation Protocol for Activation-Steering Claims, and a Cross-Family Behavioral Study of Sycophancy Under Pushback

Whether capitulation is linearly decodable from the pre-response residual stream, a prerequisite for steering-vector interventions at that locus, is asked and quantified: substring grading underestimates capitulation by 18-24 percentage points.

Saad Aamir, Muhammad Awais Bin Adil · 0 citations