INDEX

Explanations

vulgar or offensive language and terms

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_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

HCR

-0.73

 contracted

-0.68

 distingu

-0.66

âĢ¢âĢ¢âĢ¢âĢ¢

-0.66

 forfeiture

-0.64

åĬ

-0.63

Dialogue

-0.63

APH

-0.63

AUT

-0.63

 livest

-0.62

POSITIVE LOGITS

glers

1.19

bags

1.14

bag

1.07

boy

1.07

holes

1.03

heads

1.02

boys

1.00

ery

0.99

tails

0.94

hole

0.94

Activations Density 3.904%

vulgar or offensive language and terms

No Comments

No Known Activations