INDEX

Explanations

words related to the United States or its abbreviation 'US'

oai_token-act-pair · gpt-3.5-turbo

mentions of the United States (US)

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

issance

-0.77

tons

-0.74

oding

-0.71

phies

-0.70

ppings

-0.69

sbm

-0.69

istically

-0.68

osures

-0.66

aths

-0.65

ezvous

-0.63

POSITIVE LOGITS

AF

1.08

ADA

0.99

AGE

0.96

ION

0.89

MAN

0.87

SO

0.87

KER

0.86

NI

0.85

ESCO

0.84

HER

0.82

Activations Density 0.021%

words related to the United States or its abbreviation 'US'

mentions of the United States (US)

No Comments

No Known Activations