INDEX

Explanations

specific portions of a larger entity or system

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

anus

-0.78

ilial

-0.74

terday

-0.72

hesive

-0.66

clips

-0.65

ilver

-0.65

swick

-0.63

ruary

-0.63

 conclud

-0.62

onut

-0.62

POSITIVE LOGITS

of

0.62

ipation

0.61

of

0.60

aders

0.59

door

0.58

 individuals

0.57

Beh

0.57

Copyright

0.57

 exercised

0.56

ribution

0.54

Activations Density 0.024%

specific portions of a larger entity or system

No Comments

No Known Activations