INDEX

Explanations

words or phrases related to an epiphany or revelation

oai_token-act-pair · gpt-3.5-turbo

the word "ep," indicating references to episodes or elements in a series

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

 Royale

-0.82

 Breed

-0.81

haus

-0.80

ppo

-0.80

veyard

-0.79

hari

-0.78

 Grassley

-0.78

@#&

-0.77

Tonight

-0.77

 Saskatchewan

-0.77

POSITIVE LOGITS

onymous

1.12

onym

1.03

ographic

0.86

acer

0.81

ocl

0.81

iscopal

0.80

olic

0.79

¥ŀ

0.78

ically

0.75

ocry

0.75

Activations Density 0.029%

words or phrases related to an epiphany or revelation

the word "ep," indicating references to episodes or elements in a series

No Comments

No Known Activations