thinker:yingqiao-wangYingQiao Wang
Authored papers (1)
In-context learning and activation steering are the same operation on a latent-concept belief: context adds log-likelihood evidence while a steering vector adds a constant shift to the log prior odds, so the two sum in log-odds space. A four-parameter closed-form model (α, γ, a, b), log o(c|x) = b + γN^(1−α) + a·m, predicts Llama-3.1-8B-Instruct behavior on held-out steering magnitudes at r = 0.98 averaged over five persona datasets (Machiavellianism, Psychopathy, Narcissism, and two moral-nihilism sets, from Perez et al. 2022). It also predicts the crossover shot count N* at r = 0.97. Fitting is by L-BFGS with 10-fold cross-validation over steering magnitude, using Contrastive Activation Addition vectors applied at layer 12. The same model holds for Qwen-2.5-7B and Gemma-2-9B (N* r = 0.91 and 0.97), Llama-3.1-70B (r = 0.98), and a flipped-label financial sentiment task (r = 0.96). Posterior belief is sigmoidal in N^(1−α), which explains the sudden many-shot jailbreaking curves of Anil et al. (2024). Steering moves the curve left or right, and the joint (shots × magnitude) space splits into sharp phases whose boundary is computable. Steering is linear in log-odds only within a bounded range of m, beyond which behavior collapses to chance. The result implies that prompting and steering are interchangeable levers on one belief state. Small changes in either control can therefore flip behavior abruptly, and N* can be forecast for safety purposes.
More papers — OpenAlex / S2
Affiliations (1)
- Department of Psychology, Harvard University(institute)
Co-authors (9)
- Daniel Wurgaft9 shared
- Ekdeep Singh Lubana9 shared
- Hidenori Tanaka9 shared
- Eric Bigelow6 shared
- Noah D. Goodman6 shared
- Tomer David Ullman6 shared
- Eric Bigelow3 shared
- Noah Goodman3 shared
- Tomer Ullman3 shared
Their work is cited by (1)
Other inbound relations (1)
Recent mentions (1)
- papers-typedbigelow-2025-belief-dynamics.md