INDEX

Explanations

anything related to titles mentioned in the text

oai_token-act-pair · gpt-3.5-turbo

references to titles of works, such as books, movies, or articles

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

gm

-0.86

obb

-0.77

ombo

-0.75

iggs

-0.71

ierrez

-0.67

agan

-0.66

olin

-0.65

¬¼

-0.62

urrencies

-0.62

 interf

-0.62

POSITIVE LOGITS

manship

0.95

title

0.94

marks

0.84

eous

0.84

itatively

0.82

 titles

0.80

mark

0.80

 title

0.79

åĭ

0.78

ãĥĩ

0.76

Activations Density 0.017%

anything related to titles mentioned in the text

references to titles of works, such as books, movies, or articles

No Comments

No Known Activations