INDEX

Explanations

locations and directions

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

Negative Logits

bidden

-0.63

 coron

-0.56

legates

-0.53

ificant

-0.53

ihadi

-0.50

unin

-0.49

was

-0.49

translation

-0.48

pread

-0.47

anamo

-0.47

POSITIVE LOGITS

 yourself

1.05

 yourselves

0.94

 ASAP

0.92

:)

0.89

;)

0.87

 wisely

0.86

 preferably

0.85

 BEFORE

0.83

 Yourself

0.80

Activations Density 0.665%

locations and directions

No Known Activations