INDEX

Explanations

numerical percentages

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_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

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

DragonMagazine

-1.07

orc

-0.78

soDeliveryDate

-0.73

ãĥ¯ãĥ³

-0.70

ItemTracker

-0.66

Thumbnail

-0.66

rer

-0.65

actionDate

-0.63

vet

-0.63

furt

-0.62

POSITIVE LOGITS

 lashes

0.69

%);

0.63

nesia

0.61

sts

0.59

%;

0.58

 hairs

0.58

anta

0.57

yrus

0.56

 neurons

0.55

Activations Density 0.040%

numerical percentages

No Comments

No Known Activations