INDEX

Explanations

grateful

np_max-act · gemini-2.0-flash

This neuron never activates—it appears to be essentially “dead” and doesn’t detect any pattern in the text.

oai_token-act-pair · o4-mini Triggered by @xinyanhu8

New Auto-Interp

Configuration

andyrdt/saes-llama-3.1-8b-instruct/resid_post_layer_11/trainer_1

Dataset (Dashboard)

Various

Features

131,072

Data Type

float32

Hook Name

blocks.11.hook_resid_post

Architecture

standard

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 sophistic

-0.07

 Hond

-0.07

 OLED

-0.07

Ole

-0.07

 Bonnie

-0.07

ประโย

-0.07

mse

-0.06

 finish

-0.06

_ylim

-0.06

POSITIVE LOGITS

 grateful

0.10

 gratitude

0.09

\:

0.07

 thankful

0.07

 thankfully

0.07

ectar

0.06

.har

0.06

’util

0.06

 QMainWindow

0.06

バー

0.06

Activations Density 0.006%

grateful

This neuron never activates—it appears to be essentially “dead” and doesn’t detect any pattern in the text.

No Comments

No Known Activations