INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

DragonMagazine

-0.74

 mould

-0.71

 Architect

-0.60

 commercially

-0.59

 moth

-0.58

 playbook

-0.58

 fluorescent

-0.58

 communicates

-0.58

 adaptations

-0.58

 homelessness

-0.57

POSITIVE LOGITS

ciation

0.75

estate

0.73

angler

0.70

arton

0.70

lap

0.67

chain

0.67

VK

0.66

icion

0.65

itures

0.63

avage

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.