INDEX

Explanations

phrases related to the name "Beth"

oai_token-act-pair · gpt-3.5-turbo

references to specific individuals and organizations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 Erdogan

-0.79

 coup

-0.78

 Marcos

-0.77

chenko

-0.76

usk

-0.74

 Xiaomi

-0.73

 Haku

-0.72

 Martial

-0.72

 criminality

-0.72

inho

-0.71

POSITIVE LOGITS

 Beth

2.56

PBS

1.62

NIH

1.48

WB

1.24

 Martha

1.21

 panc

1.11

NES

1.10

 defund

1.05

AA

1.04

 Bard

1.01

Activations Density 0.034%

phrases related to the name "Beth"

references to specific individuals and organizations

No Comments

No Known Activations