Definition. PPO clipped surrogate [schulman2017proximal, Section 3, equation (7)] [ftip-003F]

For a clipping parameter \(\epsilon _{\mathrm {clip}}\in (0,1)\), define \(\operatorname {clip}_{\epsilon _{\mathrm {clip}}}(u) =\min (1+\epsilon _{\mathrm {clip}},\max (1-\epsilon _{\mathrm {clip}},u))\). The PPO clipped surrogate is

\[ L^{\mathrm {CLIP}}(\theta ) =\widehat {\mathbb E}_t\left [ \min \left ( \rho _t(\theta )\widehat A_t, \operatorname {clip}_{\epsilon _{\mathrm {clip}}}(\rho _t(\theta )) \widehat A_t \right )\right ]. \]

Clipping changes the sampled surrogate when the ratio moves outside the declared interval in a direction favored by the estimated advantage. It is not a hard bound on the KL divergence of the complete updated policy.