INDEX

Explanations

phrases indicating features, qualities, or items in a list

oai_token-act-pair · gpt-3.5-turbo

phrases that list or enumerate examples or factors

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

wan

-0.69

enary

-0.69

orem

-0.68

orse

-0.66

athing

-0.65

mit

-0.60

uers

-0.60

uni

-0.60

orship

-0.59

idates

-0.59

POSITIVE LOGITS

 namely

0.89

 Firstly

0.72

 notably

0.63

xual

0.62

 including

0.59

viz

0.57

etsk

0.56

 redund

0.53

 includ

0.53

 weddings

0.53

Activations Density 0.266%

phrases indicating features, qualities, or items in a list

phrases that list or enumerate examples or factors

No Comments

No Known Activations