INDEX

Explanations

proper nouns

oai_token-act-pair · gpt-3.5-turbo

the name "Aaron"

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 underwater

-0.74

 Exchange

-0.66

LSD

-0.65

 reflective

-0.64

 dump

-0.64

 Interstate

-0.63

 Islamic

-0.63

 treated

-0.62

MS

-0.62

 Folk

-0.62

POSITIVE LOGITS

aron

4.56

annon

1.24

 Baron

1.19

oris

1.13

arah

1.13

resa

1.12

ron

1.09

aro

1.09

aryn

1.08

amar

1.06

Activations Density 0.014%

proper nouns

the name "Aaron"

No Comments

No Known Activations