INDEX

Explanations

elements related to nodes and edges in a structured diagram or graph

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

RTCK

-0.40

missions

-0.37

 veiks

-0.35

 neceff

-0.34

 gydy

-0.34

########.

-0.34

égard

-0.34

 atsi

-0.33

addContainerGap

-0.33

MENAFN

-0.33

POSITIVE LOGITS

tikzpicture

0.76

 ModelExpression

0.60

Abit

0.49

TestFixture

0.48

 diagram

0.46

 ORGANISM

0.46

 papers

0.46

usepackage

0.45

isfile

0.45

LabelTagHelper

0.44

Activations Density 0.340%

elements related to nodes and edges in a structured diagram or graph

No Comments

No Known Activations