INDEX

Explanations

mentions of the word "Doge" or variations of it

oai_token-act-pair · gpt-3.5-turbo

variations of the word "dog."

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

 srfAttach

-0.68

 Torment

-0.66

Tru

-0.65

 Arri

-0.64

 Rivals

-0.64

 Hallow

-0.61

Ĥª

-0.61

urers

-0.59

furt

-0.57

catentry

-0.56

POSITIVE LOGITS

mouth

0.70

imus

0.67

iman

0.67

cially

0.63

opolis

0.63

 sake

0.62

Khe

0.61

sworth

0.60

ingham

0.60

ithmetic

0.59

Activations Density 0.140%

mentions of the word "Doge" or variations of it

variations of the word "dog."

No Comments

No Known Activations