INDEX

Explanations

content related to technical instructions or course material

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_fs3072-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

3,072

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

 marrow

-0.57

 seism

-0.54

 seismic

-0.50

Bol

-0.49

 BlackBerry

-0.49

 mull

-0.49

 Belfast

-0.49

 Hait

-0.49

 Sochi

-0.48

 critiques

-0.48

POSITIVE LOGITS

ardless

0.84

etheless

0.78

SourceFile

0.73

retty

0.67

ogether

0.66

FORE

0.66

xxxx

0.64

ignty

0.63

ucer

0.62

initely

0.62

Activations Density 6.193%

content related to technical instructions or course material

No Comments

No Known Activations