INDEX

Explanations

terms related to pornography

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs12288-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_12288

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

12,288

Data Type

torch.float32

Hook Point

blocks.8.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

 MacArthur

-0.75

pring

-0.75

 defe

-0.69

 Brewer

-0.64

soType

-0.64

 reins

-0.64

 peas

-0.63

å§«

-0.63

dur

-0.63

ees

-0.62

POSITIVE LOGITS

ographers

1.32

ography

1.15

ographer

1.13

hub

1.04

ographic

1.01

ographically

1.00

 pornography

0.95

 DVDs

0.94

 pornographic

0.94

ographies

0.91

Activations Density 0.095%

terms related to pornography

No Comments

No Known Activations