INDEX

Explanations

phrases related to political entities or ideologies

oai_token-act-pair · gpt-3.5-turbo

references to late-night shows and political left/right distinctions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

oused

-0.72

iosyncr

-0.72

è¦ļéĨĴ

-0.70

ounter

-0.68

 externalToEVAOnly

-0.67

ibly

-0.66

äºĶ

-0.66

icable

-0.66

ILY

-0.65

BILITY

-0.65

POSITIVE LOGITS

 Thing

0.97

 Ones

0.95

 Definition

0.92

 Order

0.88

Day

0.88

 Responsibility

0.88

 Works

0.87

 Roads

0.87

 Lives

0.85

 Guys

0.85

Activations Density 0.150%

phrases related to political entities or ideologies

references to late-night shows and political left/right distinctions

No Comments

No Known Activations