3 ms·
that's not really how eigenvalues work... they specifically also model the case where the result keeps changing exponentially.
by ajkjk 28d ago
that's not really how eigenvalues work... they specifically also model the case where the result keeps changing exponentially.
- mjburgess 28d agoThe claim is that the RSI operation is just finding a fixed point of improvement, RSI(LLM) = RSI(LLM) -- for an optimal LLM* which is a fixed point of RSI As for eigenvalues/vectors, they're fixed points of (1/val)A or A*val
- ekidd 28d agoEigenvectors represent fixed directions, not fixed magnitudes. From Wikipedia: > More precisely, an eigenvector v of a linear transformation T is scaled by a constant factor lambda when the linear transformation is applied to it: Tv = lambda v . In other words, repeated multiplication of an eigenvector by a matrix can still create exponential growth.
- mjburgess 26d ago> of (1/val)A