INDEX

Explanations

phrases that indicate similarity or sameness across different subjects or contexts

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Configuration

google/gemma-scope-9b-pt-res/layer_20/width_131k/average_l0_114

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.20.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 innym

-0.37

protetor

-0.36

 Einfluß

-0.36

 młod

-0.35

emploi

-0.35

occuper

-0.34

 Grüsse

-0.33

accompag

-0.33

intervento

-0.33

 nying

-0.33

POSITIVE LOGITS

 ویکی‌پدی

0.65

amoan

0.56

featureID

0.55



0.53

RTEE

0.52

 одина

0.52

 alike

0.51

исленность

0.51

も同じ

0.51

 homogenous

0.51

Activations Density 0.050%

phrases that indicate similarity or sameness across different subjects or contexts

No Comments

No Known Activations