INDEX

Explanations

locations using north, south, east, and west directional indicators

oai_token-act-pair · gpt-3.5-turbo

references to geographical directions or regions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

 Drill

-0.81

ulhu

-0.73

 Instrument

-0.71

 Eternity

-0.71

 Calculator

-0.68

 Swordsman

-0.67

 Soda

-0.66

iants

-0.63

 Instruction

-0.63

 Logo

-0.62

POSITIVE LOGITS

west

1.31

east

1.23

facing

1.12

south

1.10

western

0.98

side

0.92

eyed

0.92

earth

0.89

quartered

0.89

European

0.88

Activations Density 0.043%

locations using north, south, east, and west directional indicators

references to geographical directions or regions

No Comments

No Known Activations