INDEX

Explanations

misconceptions

np_max-act · gemini-2.0-flash

This neuron activates most strongly on occurrences of the word “misconceptions” (and its subparts) whenever text is labeling or discussing common misunderstandings.

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 sign

-0.07

comed

-0.07

 Patty

-0.07

_vals

-0.06

 gesture

-0.06

Begin

-0.06

-eye

-0.06

яття

-0.06

 relieved

-0.06

tax

-0.06

POSITIVE LOGITS

Lip

0.07

_GU

0.07

 جامع

0.06

uy�

0.06

.Filters

0.06

sic

0.06

込み

0.06

च

0.06

Bü

0.06

iosper

0.06

Activations Density 0.038%

misconceptions

This neuron activates most strongly on occurrences of the word “misconceptions” (and its subparts) whenever text is labeling or discussing common misunderstandings.

No Comments

No Known Activations