INDEX

Explanations

phrases related to people's names followed by the suffix '-endorf'

oai_token-act-pair · gpt-3.5-turbo

references to specific individuals or names associated with notable actions or events

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

 Primordial

-0.74

keyes

-0.68

Cath

-0.67

roman

-0.66

Catholic

-0.64

 trenches

-0.62

 monks

-0.60

 Catholics

-0.60

osterone

-0.59

 Tant

-0.59

POSITIVE LOGITS

endor

1.13

ffer

0.94

ning

0.87

ff

0.87

nings

0.84

lance

0.79

DAQ

0.78

ODY

0.74

lette

0.74

ption

0.73

Activations Density 0.019%

phrases related to people's names followed by the suffix '-endorf'

references to specific individuals or names associated with notable actions or events

No Comments

No Known Activations