INDEX

Explanations

mentions of siblings or related family terms

oai_token-act-pair · gpt-3.5-turbo

references to siblings and family relationships

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

protected

-0.69

vasive

-0.67

usc

-0.65

iled

-0.64

industrial

-0.63

inent

-0.62

Paris

-0.62

tical

-0.60

 Gore

-0.60

rophe

-0.60

POSITIVE LOGITS

 siblings

1.34

 sibling

0.93

 adolesc

0.87

iblings

0.86

heses

0.85

 nodd

0.81

ilial

0.81

 twins

0.81

ystem

0.81

 cousins

0.81

Activations Density 0.011%

mentions of siblings or related family terms

references to siblings and family relationships

No Comments

No Known Activations