INDEX

Explanations

words related to violent conflicts, rescue operations, and political discussions or debates

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

BOOK

-0.31

strings

-0.26

galitarian

-0.25

ertodd

-0.24

hent

-0.24

Console

-0.24

Plug

-0.24

ership

-0.24

 Eater

-0.23

WR

-0.23

POSITIVE LOGITS

aja

0.31

iatrics

0.28

deen

0.28

ãĥ¼ãĤ¯

0.28

=-=-=-=-=-=-=-=-

0.28

izons

0.27

Ã°

0.27

inder

0.27

zynski

0.27

geon

0.27

Activations Density 11.628%

words related to violent conflicts, rescue operations, and political discussions or debates

No Comments

No Known Activations