INDEX

Explanations

terms related to magnification or things that are large in scale

oai_token-act-pair · gpt-3.5-turbo

instances of the word "magn" or its variations related to magnitude or 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

RT

-0.77

BOOK

-0.75

thia

-0.75

starter

-0.73

OPLE

-0.72

 Turing

-0.71

side

-0.68

pta

-0.68

 Bleach

-0.67

ther

-0.66

POSITIVE LOGITS

 magn

1.26

ificent

1.02

ific

0.99

 tremend

0.98

olia

0.98

 magnification

0.97

 glim

0.93

itudes

0.90

animous

0.90

 eleph

0.84

Activations Density 0.007%

terms related to magnification or things that are large in scale

instances of the word "magn" or its variations related to magnitude or size

No Comments

No Known Activations