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

 ãĤµãĥ¼ãĥĨãĤ£ãĥ¯ãĥ³

-0.75

tsky

-0.73

²¾

-0.72

¶ħ

-0.71

Ļ½

-0.70

 stripes

-0.70

ī

-0.69

 hiber

-0.69

 endif

-0.69

iyah

-0.67

POSITIVE LOGITS

 Times

1.65

Times

1.26

 DMCA

0.84

OTA

0.81

 Telegraph

0.79

NYT

0.76

News

0.74

 Downing

0.72

 Registrar

0.72

times

0.71

Activations Density 0.000%

No Known Activations

This feature has no known activations.