INDEX

Explanations

references to information and knowledge

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_fs768-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_768

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

768

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

berus

-0.74

otion

-0.73

ItemTracker

-0.69

isco

-0.68

avored

-0.66

erva

-0.65

inse

-0.64

acco

-0.63

phrine

-0.63

eries

-0.62

POSITIVE LOGITS

ledge

1.25

how

1.18

ledged

1.12

lege

1.00

 firsthand

0.99

 beforehand

0.96

how

0.92

HOW

0.89

why

0.88

 whats

0.88

Activations Density 3.165%

references to information and knowledge

No Comments

No Known Activations