INDEX

Explanations

words or phrases related to motion or action

oai_token-act-pair · gpt-3.5-turbo

the use of verbs indicating movement or action

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 validity

-0.68

heed

-0.67

stice

-0.64

aspers

-0.63

essa

-0.62

Lat

-0.62

 likelihood

-0.61

inelli

-0.61

 ---------

-0.60

idable

-0.59

POSITIVE LOGITS

 nuts

1.28

 overboard

1.24

 insane

1.21

 crazy

1.20

 ballistic

1.14

mad

1.07

 bananas

1.03

 limp

1.03

ggle

1.00

 numb

1.00

Activations Density 0.065%

words or phrases related to motion or action

the use of verbs indicating movement or action

No Comments

No Known Activations