INDEX

Explanations

proper nouns and social media handles

oai_token-act-pair · gpt-3.5-turbo

instances of the abbreviation "TO."

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

 McKin

-0.70

busters

-0.69

 trademarks

-0.67

stable

-0.66

ifiers

-0.64

minster

-0.64

options

-0.64

wagen

-0.63

geist

-0.62

lain

-0.61

POSITIVE LOGITS

KEN

1.07

OME

1.04

FU

1.01

ffee

0.96

OL

0.95

ilet

0.94

OF

0.92

OOL

0.90

CLASSIFIED

0.88

YA

0.88

Activations Density 0.021%

proper nouns and social media handles

instances of the abbreviation "TO."

No Comments

No Known Activations