INDEX

Explanations

locations and directions

oai_token-act-pair · gpt-3.5-turbo

directional or locational indicators

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

icer

-0.64

zan

-0.58

aceae

-0.58

entle

-0.54

erto

-0.54

ASED

-0.53

 moderator

-0.52

 dads

-0.52

inished

-0.51

 defer

-0.50

POSITIVE LOGITS

of

1.16

ward

1.08

wards

1.06

bound

0.98

west

0.97

of

0.91

outheast

0.87

 west

0.85

 east

0.84

heast

0.80

Activations Density 0.039%

locations and directions

directional or locational indicators

No Comments

No Known Activations