INDEX

Explanations

phrases indicating importance, significance, or consequence

oai_token-act-pair · gpt-3.5-turbo

phrases emphasizing conditional meaning and significance

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

 Sabha

-0.63

 Myth

-0.63

cms

-0.62

 sugg

-0.62

YP

-0.61

osterone

-0.59

iated

-0.59

YC

-0.59

aqu

-0.58

osponsors

-0.57

POSITIVE LOGITS

ortium

0.80

thood

0.70

TRY

0.67

uni

0.64

neutral

0.63

rican

0.63

ëĭ

0.62

 grabbing

0.62

ifax

0.60

laden

0.59

Activations Density 0.091%

phrases indicating importance, significance, or consequence

phrases emphasizing conditional meaning and significance

No Comments

No Known Activations