INDEX

Explanations

the word "Wade"

oai_token-act-pair · gpt-3.5-turbo

mentions of the name "Wade."

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

 narc

-0.81

nar

-0.77

metadata

-0.75

Nar

-0.72

 ident

-0.71

------------------------------------------------

-0.70

IENT

-0.69

 meters

-0.68

orph

-0.66

aud

-0.66

POSITIVE LOGITS

 Wade

3.89

 Bosh

1.48

 Wyatt

1.30

 Deadpool

1.26

Dw

1.22

 Deng

1.19

 Dwight

1.17

Wilson

1.15

 Shane

1.13

 Shelley

1.10

Activations Density 0.018%

the word "Wade"

mentions of the name "Wade."

No Comments

No Known Activations