INDEX

Explanations

mentions of specific events or incidents involving controversy or conflict

oai_token-act-pair · gpt-3.5-turbo

references to significant social or political events

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 Subtle

-0.78

 indeed

-0.76

ETHOD

-0.74

 partName

-0.69

 moreover

-0.68

theless

-0.68

prisingly

-0.67

 anecd

-0.66

 furthermore

-0.66

 besides

-0.66

POSITIVE LOGITS

 Guantanamo

1.00

 AIDS

1.00

 Ebola

0.99

 Occupy

0.98

 Charlottesville

0.97

gay

0.97

 Benghazi

0.95

 Obamacare

0.95

 abortion

0.95

 Scientology

0.95

Activations Density 0.948%

mentions of specific events or incidents involving controversy or conflict

references to significant social or political events

No Comments

No Known Activations