INDEX

Explanations

words related to a specific name or person

oai_token-act-pair · gpt-3.5-turbo

references to specific names or individuals

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

iov

-0.76

scl

-0.72

pite

-0.68

;;;;;;;;;;;;

-0.68

 Clown

-0.66

cli

-0.65

lycer

-0.64

imens

-0.62

 Mask

-0.62

itions

-0.62

POSITIVE LOGITS

Adams

3.20

Allen

1.35

Bee

1.14

aquin

0.85

Anderson

0.80

 Beaver

0.79

 attachments

0.78

Arthur

0.77

avers

0.77

England

0.75

Activations Density 0.019%

words related to a specific name or person

references to specific names or individuals

No Comments

No Known Activations