INDEX

Explanations

simple and straightforward language, focusing on technology-related concepts and implementations

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

psey

-0.89

heimer

-0.57

 Normandy

-0.57

 Whale

-0.54

outheast

-0.53

alos

-0.52

azeera

-0.52

hner

-0.51

iannopoulos

-0.51

reon

-0.51

POSITIVE LOGITS

 arithmetic

0.69

fixes

0.60

ItemImage

0.55

tons

0.53

execute

0.53

Obj

0.52

 simple

0.51

iHUD

0.51

 isEnabled

0.50

 Etsy

0.50

Activations Density 21.287%

simple and straightforward language, focusing on technology-related concepts and implementations

No Comments

No Known Activations