INDEX

Explanations

detailed descriptions or explanations within texts

oai_token-act-pair · gpt-3.5-turbo

documents or sections that contain legal or formal assertions and summaries

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

 Sabha

-0.70

asus

-0.66

plane

-0.65

ded

-0.64

lich

-0.64

 Libre

-0.62

rip

-0.60

BUG

-0.60

 Shooter

-0.59

 Syndicate

-0.59

POSITIVE LOGITS

mith

1.40

uggest

1.39

poons

1.20

paces

1.15

pring

1.14

ettings

1.14

creen

1.12

hip

1.09

hops

1.07

hips

1.05

Activations Density 0.216%

detailed descriptions or explanations within texts

documents or sections that contain legal or formal assertions and summaries

No Comments

No Known Activations