INDEX

Explanations

names of political or official positions and titles

oai_token-act-pair · gpt-3.5-turbo

titles and roles, particularly related to authority and profession

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

Michigan

-0.77

IELD

-0.76

 brunch

-0.75

Ohio

-0.74

Wisconsin

-0.73

ilogy

-0.73

lake

-0.72

Texas

-0.72

Disney

-0.71

FOX

-0.69

POSITIVE LOGITS

 Mohammad

1.46

 Mahmoud

1.44

Vij

1.42

 Mahm

1.42

 Ibrahim

1.37

 Nikol

1.36

 Khalid

1.36

Vik

1.36

Nur

1.36

Abd

1.35

Activations Density 0.345%

names of political or official positions and titles

titles and roles, particularly related to authority and profession

No Comments

No Known Activations