INDEX

Explanations

references to interpretation or explanation of content

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

tagHelperRunner

-0.77

كويكب

-0.69

RTSC

-0.61

锈钢

-0.60

queryInterface

-0.59

SharedDtor

-0.59

 Италијани

-0.59

principalColumn

-0.59

 jspb

-0.58

enumii

-0.58

POSITIVE LOGITS

 scope

0.46

察

0.41

 pertinent

0.40

 here

0.40

 good

0.38

⤵

0.38

scope

0.38

 stack

0.38

ta

0.38

cul

0.38

Activations Density 0.035%

references to interpretation or explanation of content

No Comments

No Known Activations