INDEX

Explanations

terms related to interactive features, specifically the ability to click to enlarge something

oai_token-act-pair · gpt-3.5-turbo

words related to expanding or increasing size

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

oil

-0.72

carry

-0.66

lying

-0.65

olicited

-0.63

pron

-0.62

é¾

-0.62

spoken

-0.60

armor

-0.60

RY

-0.59

gil

-0.59

POSITIVE LOGITS

 enlarg

1.26

 enlarge

1.22

 enlarged

0.94

xual

0.92

ments

0.85

itized

0.85

larg

0.81

ement

0.79

 "$:/

0.78

ements

0.78

Activations Density 0.011%

terms related to interactive features, specifically the ability to click to enlarge something

words related to expanding or increasing size

No Comments

No Known Activations