INDEX

Explanations

the word "apt" with a high activation value

oai_token-act-pair · gpt-3.5-turbo

technical terms and jargon related to software package management and installation

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

FORE

-0.84

EGIN

-0.78

 Spears

-0.73

BIP

-0.71

OURCE

-0.69

 Zimbabwe

-0.69

 Sterling

-0.68

Creating

-0.66

Flor

-0.66

FE

-0.65

POSITIVE LOGITS

itude

1.24

apt

1.07

urous

1.03

rax

1.03

entimes

1.01

icter

1.00

icultural

0.97

omore

0.94

heon

0.93

ogether

0.93

Activations Density 0.006%

the word "apt" with a high activation value

technical terms and jargon related to software package management and installation

No Comments

No Known Activations