INDEX

Explanations

references to academic sources and authors, likely related to research or studies

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

undai

-0.90

agra

-0.79

etooth

-0.76

ison

-0.76

yon

-0.75

arlane

-0.74

xon

-0.74

rower

-0.74

yan

-0.73

apsed

-0.72

POSITIVE LOGITS

 Definitive

1.03

 Reloaded

0.97

 Volume

0.93

 Lessons

0.92

 Moving

0.91

How

0.89

 Chains

0.88

 Principles

0.85

Unt

0.85

 Reborn

0.85

Activations Density 8.040%

references to academic sources and authors, likely related to research or studies

No Comments

No Known Activations