INDEX

Explanations

references to being haunted or haunting experiences

oai_token-act-pair · gpt-3.5-turbo

mentions of haunted or ghostly themes

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

iyah

-0.77

cellence

-0.74

abet

-0.70

enture

-0.70

ouf

-0.69

reme

-0.68

iannopoulos

-0.67

erity

-0.67

utenberg

-0.66

utral

-0.66

POSITIVE LOGITS

 haunt

1.26

 haun

1.19

Haunted

1.11

 haunted

1.07

 Haunted

0.94

 haunting

0.87

crow

0.84

 attractions

0.83

 Mansion

0.80

mares

0.75

Activations Density 0.015%

references to being haunted or haunting experiences

mentions of haunted or ghostly themes

No Comments

No Known Activations