INDEX
Explanations
large numbers
The neuron fires on numeric quantity mentions (e.g. “million,” “billion,” “rows,” “500,” etc.) indicating data‐size or count metrics.
New Auto-Interp
Negative Logits
licensed
-0.07
CW
-0.06
Jwt
-0.06
гр
-0.06
employee
-0.06
ký
-0.06
くら
-0.06
glaciers
-0.06
PG
-0.06
.key
-0.06
POSITIVE LOGITS
)}}
0.08
])↵↵
0.07
.stringValue
0.07
(socket
0.07
expos
0.07
]=
0.07
])+
0.07
),"
0.07
)—
0.07
())/
0.06
Activations Density 0.016%