INDEX

Explanations

phrases related to expert opinions or analyses

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

•Layer 6 UMAP region: Mixed in middle - local

No Comments

Negative Logits

[|

-0.62

 Agric

-0.61

 Shades

-0.60

Bom

-0.57

strip

-0.56

 Interstellar

-0.56

AAAA

-0.55

fw

-0.55

reath

-0.54

 Mechdragon

-0.52

POSITIVE LOGITS

sonian

0.86

 consulted

0.84

udic

0.73

ulously

0.72

inent

0.68

 advise

0.68

ise

0.67

 diagnose

0.67

 opinion

0.66

icion

0.65

Activations Density 13.883%

phrases related to expert opinions or analyses

No Comments

No Known Activations