INDEX

Explanations

words related to military or conflict zones

oai_token-act-pair · gpt-3.5-turbo

terms related to military or governmental control and structures

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

FACE

-0.74

perature

-0.68

 Vive

-0.63

 FACE

-0.62

swer

-0.61

ker

-0.60

BOOK

-0.59

 Fair

-0.59

 Sense

-0.58

 hearty

-0.58

POSITIVE LOGITS

ament

1.14

ilit

1.09

uary

1.01

imore

1.00

hement

0.97

espie

0.94

atem

0.93

assed

0.92

ensable

0.91

assing

0.89

Activations Density 0.011%

words related to military or conflict zones

terms related to military or governmental control and structures

No Comments

No Known Activations