INDEX

Explanations

the name "Walter" in various contexts

oai_token-act-pair · gpt-3.5-turbo

mentions of the name "Walter."

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

âĶ

-0.73

 conj

-0.69

](

-0.68

 ãĥ

-0.66

 trig

-0.66

 spice

-0.65

CLUD

-0.64

 curry

-0.64

EXP

-0.63

ami

-0.63

POSITIVE LOGITS

 Walter

3.53

Wal

1.91

 Walt

1.48

 Walters

1.39

 Wally

1.33

 Werner

1.29

 Saul

1.27

 Klaus

1.25

 Ronald

1.23

 Willis

1.22

Activations Density 0.011%

the name "Walter" in various contexts

mentions of the name "Walter."

No Comments

No Known Activations