INDEX

Explanations

phrases expressing a need for change or action due to a sense of something being wrong

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

anon

-0.78

Roll

-0.72

berus

-0.72

 approvals

-0.71

lees

-0.70

 Merchants

-0.70

 Foley

-0.69

 Sprint

-0.66

 Fighters

-0.66

Dou

-0.66

POSITIVE LOGITS

 somew

0.91

else

0.87

 intangible

0.87

ificant

0.84

ython

0.83

 else

0.83

chy

0.79

angible

0.79

worldly

0.79

 somewhere

0.78

Activations Density 11.816%

phrases expressing a need for change or action due to a sense of something being wrong

No Comments

No Known Activations