INDEX

Explanations

specific surnames with a high level of activation

oai_token-act-pair · gpt-3.5-turbo

the name "Hogan" in various contexts related to debate and media

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

iterranean

-0.78

reen

-0.77

aution

-0.77

nesota

-0.73

ERAL

-0.70

PsyNetMessage

-0.69

ropolitan

-0.68

aeda

-0.68

anwhile

-0.68

ometimes

-0.68

POSITIVE LOGITS

 Hogan

1.27

 McMahon

0.72

weed

0.71

pin

0.71

loo

0.68

lings

0.67

Dug

0.67

lord

0.66

 Bros

0.65

 Suit

0.63

Activations Density 0.007%

specific surnames with a high level of activation

the name "Hogan" in various contexts related to debate and media

No Comments

No Known Activations