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

 Medic

-0.73

 bool

-0.66

)))

-0.65

dylib

-0.64

 LIMITED

-0.63

Plus

-0.62

))))

-0.62

00200000

-0.61

DoS

-0.59

 Plus

-0.58

POSITIVE LOGITS

NPR

0.82

 Berk

0.81

kj

0.71

 arsen

0.71

igham

0.71

tyard

0.70

agues

0.69

burgh

0.68

odon

0.68

pedia

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.