INDEX

Explanations

mentions of government-related terms and discussions in text

oai_token-act-pair · gpt-3.5-turbo

mentions of government and its actions or roles

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

ranging

-0.72

joining

-0.71

actionDate

-0.70

sembly

-0.68

uberty

-0.68

 Junction

-0.67

interstitial

-0.64

 newcom

-0.64

upiter

-0.64

auri

-0.63

POSITIVE LOGITS

 collects

1.14

 regulates

1.11

 prosec

1.11

 owes

1.10

 recognizes

1.10

 considers

1.10

 refuses

1.10

 deems

1.08

 knows

1.08

 believes

1.06

Activations Density 0.242%

mentions of government-related terms and discussions in text

mentions of government and its actions or roles

No Comments

No Known Activations