INDEX

Explanations

specific advertisements

oai_token-act-pair · gpt-3.5-turbo

instances of the word "Ad" as a marker related to advertising or ad content

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

Jr

-0.69

 Kubrick

-0.67

SOC

-0.66

 Stronghold

-0.66

 trout

-0.62

 Weir

-0.62

 Kafka

-0.60

PBS

-0.59

 tuberculosis

-0.59

 scrimmage

-0.59

POSITIVE LOGITS

vertis

1.46

mittedly

1.36

ept

1.32

vantage

1.31

olescent

1.30

olesc

1.27

vance

1.26

option

1.25

elaide

1.25

hesive

1.21

Activations Density 0.019%

specific advertisements

instances of the word "Ad" as a marker related to advertising or ad content

No Comments

No Known Activations