sas proc lifereg unlogged response

Question

我在数据集上执行了SAS PROC LIFEREG，假设基线分布为广义伽玛。我使用的命令是：

proc lifereg data=work;
model time*censor(0)=mqlp bsid mkd1 mkd1x  mkd2 szsd stkv turn / distribution=gamma ;
run;

我得到了合适的统计数据：

Fit Statistics
-2 Log Likelihood   4668.169
AIC (smaller is better) 4690.169
AICC (smaller is better)    4690.209
BIC (smaller is better) 4764.887

Fit Statistics (Unlogged Response)
-2 Log Likelihood   9351.562
Gamma AIC (smaller is better)   9373.562
Gamma AICC (smaller is better)  9373.602
Gamma BIC (smaller is better)   9448.280

为什么未记录的repsonse与记录的repsonse有不同的可能性？这两种情况有什么区别？