INDEX

Explanations

news headlines, especially those involving strong statements or opinions

oai_token-act-pair · gpt-3.5-turbo

instances of the word "MORE" related to various topics or issues

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

liest

-0.87

stood

-0.86

bows

-0.85

vet

-0.70

eries

-0.69

ades

-0.68

bred

-0.68

oes

-0.67

keeping

-0.66

bow

-0.65

POSITIVE LOGITS

ado

0.95

 than

0.93

 VIDEOS

0.88

enhagen

0.79

HEAD

0.78

 Than

0.77

 INFORMATION

0.75

MORE

0.75

HUD

0.72

 importantly

0.72

Activations Density 0.017%

news headlines, especially those involving strong statements or opinions

instances of the word "MORE" related to various topics or issues

No Comments

No Known Activations