INDEX

Explanations

explosive or impactful events or objects

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ousing

-0.18

naire

-0.17

ously

-0.17

resp

-0.15

aper

-0.15

CTR

-0.15

icating

-0.15

rior

-0.15

vier

-0.15

erent

-0.14

POSITIVE LOGITS

ificent

0.20

mingham

0.19

pheus

0.19

isson

0.18

andr

0.17

lasses

0.17

 Andersen

0.17

sung

0.17

wig

0.17

sonian

0.17

Activations Density 10.503%

explosive or impactful events or objects

No Comments

No Known Activations