INDEX

Explanations

political figures and their party affiliations

oai_token-act-pair · gpt-3.5-turbo

sequences of characters typically associated with legislative discussions or government-related topics

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

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

gir

-0.78

rhy

-0.77

 unconditional

-0.75

 barr

-0.71

 contempor

-0.71

 sporadic

-0.69

 rhyth

-0.69

 casual

-0.69

gou

-0.68

 attackers

-0.68

POSITIVE LOGITS

Calif

1.59

Fla

1.53

Va

1.43

Utah

1.40

Ga

1.40

Tenn

1.40

Texas

1.38

Ari

1.37

CA

1.37

TX

1.37

Activations Density 0.024%

political figures and their party affiliations

sequences of characters typically associated with legislative discussions or government-related topics

No Comments

No Known Activations