INDEX

Explanations

instances of code or comments related to computational processes or function instructions

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

colhead

-0.76

 AssemblyCulture

-0.74

InitVars

-0.73

OGND

-0.71

<unused52>

-0.69

<unused14>

-0.69

<unused41>

-0.69

<unused51>

-0.69

<unused16>

-0.69

[@BOS@]

-0.69

POSITIVE LOGITS

Counter

0.44

IfExists

0.40

VELOP

0.38

OVERR

0.38

Buenos

0.38

Sa

0.38

0.37

Activations Density 0.020%

instances of code or comments related to computational processes or function instructions

No Comments

No Known Activations