INDEX
Explanations
sequences of repeated dashes or lines in the document
New Auto-Interp
Negative Logits
?».
-0.91
']}
-0.90
"]}
-0.89
'">
-0.86
?»
-0.85
]';
-0.85
)».
-0.84
']
-0.84
!».
-0.82
();}
-0.81
POSITIVE LOGITS
----------------
2.21
---------------
1.28
--------------
1.20
--------
1.06
------------
1.04
-------------
1.01
-----------
0.99
------
0.97
---------
0.92
----------
0.89
Activations Density 0.220%