INDEX

Explanations

the word "Dennis" and related variations

oai_token-act-pair · gpt-3.5-turbo

repeated mentions of the name "Dennis" in various contexts

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

UX

-0.68

 loot

-0.68

nav

-0.67

 leve

-0.67

val

-0.67

 sync

-0.66

 looting

-0.64

 lime

-0.64

âĸº

-0.64

alt

-0.63

POSITIVE LOGITS

 Dennis

3.41

ennis

1.97

 Darrell

1.31

 Dirk

1.30

 Denis

1.28

 Douglas

1.27

 Doug

1.25

 Curtis

1.24

 Ronnie

1.19

 Daryl

1.19

Activations Density 0.015%

the word "Dennis" and related variations

repeated mentions of the name "Dennis" in various contexts

No Comments

No Known Activations