INDEX

Explanations

information about building upon previous work or concepts

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

BUT

-0.73

hammer

-0.70

Kids

-0.67

soDeliveryDate

-0.66

Ø§Ø

-0.66

ogun

-0.65

 Penguins

-0.63

Bit

-0.63

SourceFile

-0.62

TY

-0.62

POSITIVE LOGITS

 behalf

1.05

rawdownloadcloneembedreportprint

0.82

erous

0.82

 stage

0.78

 foundations

0.78

top

0.77

 occasion

0.77

stage

0.76

 rooft

0.72

 premise

0.71

Activations Density 13.007%

information about building upon previous work or concepts

No Comments

No Known Activations