INDEX

Explanations

words related to entering a particular location or field, such as entering politics, a country, a business, or an organization

oai_token-act-pair · gpt-3.5-turbo

references to political contexts and locations

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

 Poles

-0.62

 inclusion

-0.61

idth

-0.61

 Swed

-0.57

orsi

-0.56

 commenter

-0.56

POV

-0.56

 divergence

-0.55

 Polish

-0.54

 penet

-0.54

POSITIVE LOGITS

 fray

0.88

 illegally

0.76

nas

0.72

hess

0.72

 premises

0.70

 voluntarily

0.69

wagon

0.68

illery

0.67

 kitchens

0.66

una

0.66

Activations Density 0.680%

words related to entering a particular location or field, such as entering politics, a country, a business, or an organization

references to political contexts and locations

No Comments

No Known Activations