INDEX

Explanations

phrases related to criticism or disapproval

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 assassinated

-0.72

 tampering

-0.71

mith

-0.71

bone

-0.69

 guarant

-0.68

Cru

-0.68

 Pupp

-0.67

shake

-0.66

kee

-0.66

jong

-0.66

POSITIVE LOGITS

itures

1.52

iture

1.29

 calories

1.02

folios

0.95

GBT

0.94

ewater

0.93

 resources

0.91

ptives

0.91

 energies

0.91

 ACTIONS

0.91

Activations Density 1.913%

phrases related to criticism or disapproval

No Comments

No Known Activations