INDEX

Explanations

phrases indicating importance or value

oai_token-act-pair · gpt-3.5-turbo

mentions of significance or relevance in various contexts

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

resp

-0.77

vae

-0.70

hma

-0.66

cker

-0.66

igans

-0.65

Interstitial

-0.63

nar

-0.62

 Rider

-0.62

IDES

-0.61

anus

-0.61

POSITIVE LOGITS

 significance

1.07

istically

0.80

 importance

0.80

ional

0.78

notations

0.77

xual

0.75

obyl

0.74

hift

0.74

 seals

0.73

notation

0.73

Activations Density 0.009%

phrases indicating importance or value

mentions of significance or relevance in various contexts

No Comments

No Known Activations