INDEX

Explanations

words related to expressing strong emotions or desires

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

 guiActiveUnfocused

-0.72

adena

-0.70

icist

-0.63

 Presidency

-0.61

 guiName

-0.60

YING

-0.58

 Azerbai

-0.58

Reviewer

-0.58

 Various

-0.57

Delivery

-0.57

POSITIVE LOGITS

 knew

1.06

owe

1.04

 forgot

0.93

 swore

0.89

 despise

0.88

 need

0.88

 chose

0.87

 rode

0.86

 wrote

0.86

 crave

0.85

Activations Density 11.315%

words related to expressing strong emotions or desires

No Comments

No Known Activations