INDEX

Explanations

mentions of political parties, party officials, and party activities

oai_token-act-pair · gpt-3.5-turbo

references to political parties

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

alam

-0.81

alez

-0.79

ravings

-0.76

aneously

-0.75

apons

-0.70

obar

-0.68

Hoo

-0.65

 Compton

-0.64

 Wheat

-0.63

 Grizz

-0.63

POSITIVE LOGITS

 affiliation

1.29

 affili

1.09

 leadership

1.05

 caucus

1.00

 nominate

0.98

Leader

0.97

 allegiance

0.97

 nominating

0.95

 membership

0.95

 congress

0.94

Activations Density 0.082%

mentions of political parties, party officials, and party activities

references to political parties

No Comments

No Known Activations