INDEX

Explanations

user-specific identifiers, likely from social media posts or profiles

oai_token-act-pair · gpt-3.5-turbo

letters and abbreviations used in social media handles or accounts

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

LEY

-0.78

EStream

-0.77

 BART

-0.74

NESS

-0.72

 Kers

-0.71

 Chaser

-0.69

ĪĴ

-0.69

 Shepherd

-0.65

 Mellon

-0.65

ICLE

-0.65

POSITIVE LOGITS

otor

1.01

arc

1.00

umbers

0.98

cs

0.95

ynamic

0.95

jac

0.94

cd

0.94

ounds

0.94

aint

0.93

ixels

0.93

Activations Density 0.117%

user-specific identifiers, likely from social media posts or profiles

letters and abbreviations used in social media handles or accounts

No Comments

No Known Activations