INDEX

Explanations

references to military actions and government investigations

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

IVer

-0.41

EClass

-0.39

revert

-0.39

Dur

-0.39

getMinutes

-0.38

Shine

-0.38

 Backend

-0.38

Bored

-0.37

Dt

-0.37

WithIOException

-0.37

POSITIVE LOGITS

 MainAxisSize

0.59

 Normdatei

0.52

IntoConstraints

0.51

AccessorTable

0.45

 Ähn

0.43

protoimpl

0.43

 gefährlich

0.43

 OMITBAD

0.42

 argint

0.42

HomeAsUp

0.41

Activations Density 0.001%

references to military actions and government investigations

No Comments

No Known Activations