INDEX

Explanations

references to structures like garages, houses, and dormitories

oai_token-act-pair · gpt-3.5-turbo

words related to structures and places

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

 compe

-0.73

 neut

-0.70

 streng

-0.65

Ã¡s

-0.65

SEN

-0.64

 Refer

-0.63

 Contrast

-0.62

 Moderate

-0.61

Lif

-0.61

 Childhood

-0.60

POSITIVE LOGITS

lain

1.02

hare

0.99

rats

0.93

velt

0.84

itory

0.81

rooms

0.78

pole

0.78

maid

0.76

itual

0.74

cloth

0.74

Activations Density 0.226%

references to structures like garages, houses, and dormitories

words related to structures and places

No Comments

No Known Activations