INDEX

Explanations

information related to legal court cases, police activity, and controversies

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

direct references to individuals and their statements or actions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 12-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.hook_resid_post

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.11.hook_resid_post

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

astery

-0.87

 glacier

-0.81

 cher

-0.81

 glac

-0.78

 pastry

-0.77

 Armenian

-0.77

 Grizz

-0.77

 rapp

-0.76

 Alps

-0.76

 avalanche

-0.75

POSITIVE LOGITS

Johnson

2.77

 Johnson

2.74

 Weld

1.67

john

1.36

 Libertarian

1.29

 Lyndon

1.17

LB

1.14

 john

1.06

 Johnston

1.06

 Johns

1.06

Activations Density 0.175%

information related to legal court cases, police activity, and controversies

direct references to individuals and their statements or actions

No Comments

No Known Activations