INDEX

Explanations

methods and properties related to programming classes and structures, with a focus on self-referential concepts

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Configuration

neuronpedia/gpt2-small__res_scefr-ajt/6-res_scefr-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

 GOODMAN

-1.12

 XIII

-1.10

 Syndicate

-0.98

 RELEASE

-0.97

 Ashe

-0.97

ournal

-0.95

ī

-0.94

iatus

-0.93

SHIP

-0.93

 Flavoring

-0.93

POSITIVE LOGITS

destruct

1.70

 destruct

1.31

same

1.30

ridges

1.26

ortium

1.13

 explanatory

1.12

uded

1.11

uted

1.08

ridge

1.06

contained

1.05

Activations Density 0.456%

methods and properties related to programming classes and structures, with a focus on self-referential concepts

No Comments

No Known Activations