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

ãĤ¦ãĤ¹

-0.86

Cause

-0.81

-0.78

maxwell

-0.77

inventoryQuantity

-0.73

DEV

-0.72

MpServer

-0.71

Osw

-0.71

 triangles

-0.70

":[

-0.69

POSITIVE LOGITS

 sacked

0.72

 shell

0.69

ire

0.68

 rushed

0.67

wards

0.65

 prepared

0.64

 stone

0.63

 flooded

0.63

 presses

0.63

 tunnel

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.