INDEX

Explanations

the word "slew" followed by nouns

oai_token-act-pair · gpt-3.5-turbo

phrases indicating a large quantity or a collections of items

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

abad

-0.75

ittens

-0.69

prints

-0.68

agate

-0.66

istan

-0.65

 fundament

-0.64

anium

-0.63

 Gutenberg

-0.62

 Tycoon

-0.60

 nearest

-0.59

POSITIVE LOGITS

otom

0.74

efully

0.72

dozen

0.68

mount

0.68

 array

0.67

 slew

0.67

ingly

0.63

 notable

0.61

uss

0.61

months

0.61

Activations Density 0.103%

the word "slew" followed by nouns

phrases indicating a large quantity or a collections of items

No Comments

No Known Activations