INDEX

Explanations

negative words and terms related to errors or mistakes

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

vation

-0.51

awn

-0.51

Male

-0.50

ult

-0.49

Female

-0.49

SPA

-0.49

icipated

-0.49

DragonMagazine

-0.49

grave

-0.47

-0.45

POSITIVE LOGITS

 workaround

0.59

 tricky

0.56

 terminology

0.53

 semantics

0.52

kered

0.51

 tricks

0.51

 schemes

0.50

 complicate

0.49

 complicated

0.49

 gadgets

0.49

Activations Density 15.367%

negative words and terms related to errors or mistakes

No Comments

No Known Activations