INDEX

Explanations

words related to control and authority

oai_token-act-pair · gpt-3.5-turbo

terms related to control and authority

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 partName

-0.64

ggles

-0.64

pires

-0.58

»Ĵ

-0.56

urry

-0.56

 pione

-0.56

azes

-0.54

guyen

-0.54

 whisper

-0.53

minist

-0.53

POSITIVE LOGITS

 because

0.96

 whereas

0.89

because

0.83

0.78

 despite

0.76

".

0.73

 "âĢ¦

0.71

 unfairly

0.71

 although

0.71

 regardless

0.71

Activations Density 1.340%

words related to control and authority

terms related to control and authority

No Comments

No Known Activations