INDEX

Explanations

prepositions followed by a location or a condition, most commonly "on" followed by a descriptor

oai_token-act-pair · gpt-3.5-turbo

phrases indicating actions or events occurring "on" a specific time or place

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

acters

-0.73

ocus

-0.72

binding

-0.69

emort

-0.69

INESS

-0.66

ennes

-0.65

Definition

-0.64

ï¸ı

-0.63

YY

-0.63

Ba

-0.62

POSITIVE LOGITS

 behalf

1.70

 weekends

1.16

site

1.09

 Sundays

1.08

 Saturdays

1.01

shore

1.00

 rooft

0.98

 occasion

0.96

 fumes

0.96

eday

0.89

Activations Density 0.210%

prepositions followed by a location or a condition, most commonly "on" followed by a descriptor

phrases indicating actions or events occurring "on" a specific time or place

No Comments

No Known Activations