INDEX

Explanations

phrases indicating legal or formal directives and actions, potentially related to government or law enforcement

oai_token-act-pair · gpt-3.5-turbo

symbols or special characters used in a political context

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

mares

-0.69

ousel

-0.67

 fascination

-0.67

 Elys

-0.67

 Niet

-0.65

 adore

-0.65

 Leica

-0.64

rom

-0.63

 fertility

-0.62

 Romance

-0.62

POSITIVE LOGITS

ª

1.39

Ĵ

1.30

ĳ

1.28

¤

1.24

®

1.22

«

1.22

ı

1.16

Ĳ

1.16

°

1.15

ĸ

1.13

Activations Density 0.190%

phrases indicating legal or formal directives and actions, potentially related to government or law enforcement

symbols or special characters used in a political context

No Comments

No Known Activations