INDEX

Explanations

positive or negative evaluations and attitudes towards different topics or circumstances

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 kindred

-0.55

unta

-0.52

adr

-0.52

him

-0.49

bub

-0.47

enary

-0.47

lÃ©

-0.46

aternity

-0.46

 ãĤµãĥ¼ãĥĨãĤ£ãĥ¯ãĥ³

-0.45

fold

-0.45

POSITIVE LOGITS

 Thoughts

0.68

Person

0.56

 though

0.55

 however

0.53

 Length

0.53

 Rating

0.52

itionally

0.52

é¾įåĸļå£«

0.52

Overall

0.51

alyst

0.50

Activations Density 12.161%

positive or negative evaluations and attitudes towards different topics or circumstances

No Comments

No Known Activations