INDEX

Explanations

words related to conflict, controversy, or negative events

oai_token-act-pair · gpt-3.5-turbo

the word "inc" in various contexts, suggesting a focus on incomplete or ongoing actions or conditions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

buster

-0.79

wagen

-0.75

geist

-0.75

hyde

-0.71

Ĥ¬

-0.71

DragonMagazine

-0.70

externalActionCode

-0.69

ĸļ

-0.69

assetsadobe

-0.68

 Halls

-0.67

POSITIVE LOGITS

orporated

1.33

ursions

1.32

ursion

1.29

ision

1.27

isions

1.17

idences

1.14

arcer

1.14

itement

1.12

isive

1.10

enture

1.10

Activations Density 0.008%

words related to conflict, controversy, or negative events

the word "inc" in various contexts, suggesting a focus on incomplete or ongoing actions or conditions

No Comments

No Known Activations