INDEX

Explanations

instances of the word "some" followed by positive descriptions or actions

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

istan

-0.95

gon

-0.95

raid

-0.94

SPONSORED

-0.92

ipper

-0.92

gang

-0.90

peed

-0.88

ocene

-0.88

acus

-0.86

agents

-0.84

POSITIVE LOGITS

place

1.35

ones

1.24

 semblance

1.15

 serious

1.13

body

1.12

 consolation

1.07

 additional

1.07

how

1.06

 decent

1.02

 sort

1.01

Activations Density 0.835%

instances of the word "some" followed by positive descriptions or actions

No Comments

No Known Activations