INDEX

Explanations

action-related words such as 'walk' and 'went'

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

actions and processes related to movement or organization

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 12-res-jb

Configuration

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

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_post

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

venge

-0.60

iru

-0.59

Featured

-0.51

 pseudonym

-0.50

ilitarian

-0.50

 pronounced

-0.50

uti

-0.50

AUT

-0.49

 frivol

-0.49

][

-0.49

POSITIVE LOGITS

 through

2.17

 thru

1.86

through

1.76

 Through

1.55

Through

1.44

 trough

1.22

THR

0.88

 past

0.88

past

0.87

 around

0.87

Activations Density 0.612%

action-related words such as 'walk' and 'went'

actions and processes related to movement or organization

No Comments

No Known Activations