INDEX

Explanations

names of individuals

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

sites

-0.83

ihar

-0.73

tical

-0.71

atched

-0.68

 largeDownload

-0.67

iatrics

-0.66

cule

-0.65

cedented

-0.65

phis

-0.64

around

-0.63

POSITIVE LOGITS

 Nixon

1.03

 Dawkins

1.01

 Pryor

0.89

 Simmons

0.83

 Johnson

0.80

 Strauss

0.80

 Wagner

0.78

 Madden

0.77

 Stall

0.76

 Clarke

0.76

Activations Density 0.022%

names of individuals

No Comments

No Known Activations