INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

\":

-0.72

 Learns

-0.71

 Proceed

-0.69

 Vehicles

-0.68

yss

-0.67

Gau

-0.67

JECT

-0.67

UTE

-0.65

ä¸ī

-0.64

BAT

-0.64

POSITIVE LOGITS

oen

0.90

ife

0.79

istors

0.78

former

0.75

abi

0.74

ancial

0.74

obi

0.72

stakes

0.71

ogan

0.71

IFE

0.70

Activations Density 0.000%

No Known Activations

This feature has no known activations.