INDEX

Explanations

the word 'me' in various contexts

oai_token-act-pair · gpt-3.5-turbo

phrases requesting or asking for something

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

icion

-0.84

raviolet

-0.73

osterone

-0.72

imil

-0.71

itect

-0.69

earable

-0.66

raints

-0.66

rotein

-0.64

esity

-0.64

lycer

-0.64

POSITIVE LOGITS

adow

0.93

zzo

0.89

adows

0.88

 personally

0.88

lees

0.88

cca

0.71

imei

0.71

zz

0.70

 myself

0.66

ered

0.65

Activations Density 0.055%

the word 'me' in various contexts

phrases requesting or asking for something

No Comments

No Known Activations