INDEX

Explanations

references to legal names or words related to locations

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

aeda

-0.90

EED

-0.88

REE

-0.85

URI

-0.84

Ø§ÙĦ

-0.80

 Abyssal

-0.79

å°Ĩ

-0.78

 unauthorized

-0.75

éĹĺ

-0.75

kick

-0.74

POSITIVE LOGITS

idays

1.56

iday

1.44

ocaust

1.44

ocene

1.25

mberg

1.15

tro

1.13

stery

1.12

oho

1.10

brook

1.09

isky

1.09

Activations Density 5.851%

references to legal names or words related to locations

No Comments

No Known Activations