INDEX

Explanations

instances where something is positioned, oriented, or moved in a particular direction

oai_token-act-pair · gpt-3.5-turbo

directional references in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

BuyableInstoreAndOnline

-0.82

Politics

-0.69

 resurgence

-0.66

 resurg

-0.65

 pioneers

-0.61

ethical

-0.61

 Freak

-0.61

 cynical

-0.60

sav

-0.60

 renaissance

-0.60

POSITIVE LOGITS

 perpendicular

1.17

 horizontally

1.16

(<

1.15

 depending

1.03

 perpend

1.02

1.00

 vertically

1.00

(>

0.98

(=

0.97

 whereas

0.97

Activations Density 0.470%

instances where something is positioned, oriented, or moved in a particular direction

directional references in various contexts

No Comments

No Known Activations