INDEX

Explanations

references to the original version of something

oai_token-act-pair · gpt-3.5-turbo

references to the concept of "original" 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 @ 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

robe

-0.74

abad

-0.70

omo

-0.69

uffed

-0.69

walk

-0.68

rooms

-0.66

Across

-0.64

anship

-0.64

 Particularly

-0.63

cture

-0.63

POSITIVE LOGITS

 incarnation

1.00

 originals

0.95

 original

0.92

original

0.80

ITY

0.78

ity

0.78

 incarn

0.78

istar

0.76

 creator

0.75

 predecessor

0.75

Activations Density 0.014%

references to the original version of something

references to the concept of "original" in various contexts

No Comments

No Known Activations