INDEX

Explanations

phrases related to complex social, political, or technical concepts involving detailed explanations and processes

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Configuration

neuronpedia/gpt2-small__res_scl-ajt/6-res_scl-ajt

Prompts (Dashboard)

12,288 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

46,080

Data Type

torch.float32

Hook Point

blocks.6.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

apollo-research/Skylion007-openwebtext-tokenizer-gpt2

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

TI

-0.55

apult

-0.53

itudinal

-0.53

onew

-0.53

 ãĤµãĥ¼ãĥĨãĤ£ãĥ¯ãĥ³

-0.52

hat

-0.52

zon

-0.51

rift

-0.50

 desks

-0.48

heit

-0.48

POSITIVE LOGITS

 removing

0.56

 brute

0.56

 involving

0.55

ioned

0.54

 considerable

0.53

cius

0.53

rel

0.52

 extracting

0.52

 swapping

0.52

olved

0.51

Activations Density 10.801%

phrases related to complex social, political, or technical concepts involving detailed explanations and processes

No Comments

No Known Activations