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

itars

-0.83

cum

-0.70

 Ezek

-0.67

nikov

-0.63

CAM

-0.63

Wad

-0.61

CLR

-0.61

RAY

-0.60

igate

-0.60

 addressed

-0.59

POSITIVE LOGITS

akedown

0.73

schild

0.70

adra

0.68

OWS

0.68

onde

0.67

 refuel

0.66

¯

0.65

furt

0.64

owed

0.64

 raids

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.