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

 mathemat

-0.85

ipal

-0.74

Split

-0.72

ãĥ©ãĥ³

-0.70

ãĥ¯ãĥ³

-0.70

ãĤ°

-0.69

displayText

-0.67

ãĥ¼ãĤ¯

-0.66

Else

-0.65

rique

-0.64

POSITIVE LOGITS

arily

0.73

 grat

0.69

 joking

0.66

aw

0.66

 laughing

0.64

othe

0.63

ooo

0.63

wo

0.59

 Blade

0.58

 laugh

0.57

Activations Density 0.000%

No Known Activations

This feature has no known activations.