INDEX

Explanations

instances where strong opinions or criticisms are expressed

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

thood

-0.74

utm

-0.69

reve

-0.68

 suppose

-0.68

ciplinary

-0.66

craft

-0.66

Malley

-0.66

Iter

-0.65

FontSize

-0.64

 thereof

-0.63

POSITIVE LOGITS

 same

1.62

 hardest

1.38

 slightest

1.30

 brunt

1.25

ses

1.25

 entirety

1.21

 toughest

1.20

same

1.20

 fastest

1.13

 entire

1.13

Activations Density 0.225%

instances where strong opinions or criticisms are expressed

No Comments

No Known Activations