INDEX

Explanations

the name "Recep Tayyip Erdogan" and related mentions of world leaders

oai_token-act-pair · gpt-3.5-turbo

mentions of specific political figures, particularly Recep Tayyip Erdogan

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

birth

-0.82

psc

-0.81

Medic

-0.77

tests

-0.77

Recommend

-0.76

hospital

-0.72

advertisement

-0.72

source

-0.71

PASS

-0.69

BO

-0.69

POSITIVE LOGITS

 Tayyip

1.59

 Erdogan

1.20

Erd

1.19

 Recep

1.11

 Sheikh

0.85

Suk

0.82

 Jinping

0.81

Tay

0.78

Yus

0.78

ctuary

0.78

Activations Density 0.008%

the name "Recep Tayyip Erdogan" and related mentions of world leaders

mentions of specific political figures, particularly Recep Tayyip Erdogan

No Comments

No Known Activations