INDEX

Explanations

frequent words used by reviewers to express their preferences, such as "favorite."

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

staking

-1.29

sha

-1.04

krit

-0.99

 Centauri

-0.98

urally

-0.97

eva

-0.97

gypt

-0.96

manship

-0.96

aucas

-0.96

ucl

-0.95

POSITIVE LOGITS

 favorite

1.05

 favorites

0.89

 Likes

0.89

ãĥīãĥ©ãĤ´ãĥ³

0.87

ãĥł

0.85

 Candy

0.82

ROR

0.82

 âĺħ

0.82

 reunion

0.81

 fruits

0.81

Activations Density 0.019%

frequent words used by reviewers to express their preferences, such as "favorite."

No Comments

No Known Activations