INDEX

Explanations

words related to written content, such as books, corners, and campaigns

oai_token-act-pair · gpt-3.5-turbo

words associated with various aspects of conflict and community issues

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

 Forestry

-0.69

cius

-0.65

aph

-0.63

 Tags

-0.62

ãĤ¨ãĥ«

-0.60

afort

-0.60

ilers

-0.60

ergy

-0.60

orno

-0.59

ighed

-0.59

POSITIVE LOGITS

stros

0.77

ichick

0.65

isphere

0.64

ATURE

0.64

OULD

0.64

 extends

0.64

cture

0.64

 initiative

0.61

é»Ĵ

0.59

smanship

0.59

Activations Density 0.286%

words related to written content, such as books, corners, and campaigns

words associated with various aspects of conflict and community issues

No Comments

No Known Activations