INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 Leilan

-0.65

 ashore

-0.64

Pink

-0.61

 Mavericks

-0.58

 explorer

-0.57

 Participation

-0.57

 daily

-0.57

 colourful

-0.56

 navy

-0.56

ACS

-0.56

POSITIVE LOGITS

ideo

0.85

heim

0.78

rights

0.75

 Hollow

0.72

nob

0.72

rist

0.71

ENC

0.71

nesota

0.69

Zone

0.68

ax

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.