INDEX

Explanations

medical terms and conditions

oai_token-act-pair · gpt-3.5-turbo

terms related to being armed or having weapons

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

Gate

-0.77

MAL

-0.77

MRI

-0.75

Hub

-0.75

 Woodward

-0.73

RGB

-0.72

TX

-0.71

Â¢

-0.71

Publisher

-0.71

Bee

-0.71

POSITIVE LOGITS

 armed

1.16

 arming

0.92

 robbery

0.92

 uprising

0.91

 conflic

0.90

EStream

0.90

 revolt

0.87

 guards

0.87

 volunte

0.86

ilitary

0.86

Activations Density 0.011%

medical terms and conditions

terms related to being armed or having weapons

No Comments

No Known Activations