INDEX

Explanations

booleans, likely focusing on values 9 and 10

oai_token-act-pair · gpt-3.5-turbo

mentions of the National Rifle Association (NRA) and related terms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

lace

-0.95

acity

-0.78

 MacArthur

-0.73

tein

-0.72

 Corpus

-0.69

 ç¥ŀ

-0.69

bars

-0.67

ician

-0.66

iewicz

-0.65

terson

-0.65

POSITIVE LOGITS

RA

1.08

VE

1.04

VEN

0.97

BB

0.94

ISE

0.93

GE

0.93

BA

0.93

BO

0.92

BILITY

0.91

HAM

0.90

Activations Density 0.007%

booleans, likely focusing on values 9 and 10

mentions of the National Rifle Association (NRA) and related terms

No Comments

No Known Activations