INDEX

Explanations

words related to private or privileged matters

oai_token-act-pair · gpt-3.5-turbo

references to privacy and privileged information

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

enegger

-0.72

calling

-0.70

grass

-0.68

rers

-0.67

LER

-0.66

WAYS

-0.66

 Ducks

-0.65

ORN

-0.64

 Modest

-0.64

 snail

-0.63

POSITIVE LOGITS

ileged

1.52

ilege

1.44

acies

1.12

ately

1.10

urrent

1.00

ession

0.97

iless

0.95

ounced

0.92

 mathemat

0.92

oxy

0.92

Activations Density 0.011%

words related to private or privileged matters

references to privacy and privileged information

No Comments

No Known Activations