INDEX

Explanations

mentions of the word "Nap" in various contexts

oai_token-act-pair · gpt-3.5-turbo

references to "Napoleon" or related terms

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

 Gemini

-0.91

DonaldTrump

-0.77

UCT

-0.75

 Carbuncle

-0.70

×Ļ×

-0.70

````

-0.68

âĸ¬âĸ¬

-0.67

âĢ¢âĢ¢âĢ¢âĢ¢

-0.66

ÂłÂłÂł

-0.66

 erroneous

-0.65

POSITIVE LOGITS

oleon

1.39

olit

1.32

alm

1.31

olitan

1.28

erville

1.05

oli

0.98

rose

0.95

ixels

0.91

ole

0.89

ster

0.89

Activations Density 0.012%

mentions of the word "Nap" in various contexts

references to "Napoleon" or related terms

No Comments

No Known Activations