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

Tro

-0.71

%]

-0.68

aders

-0.67

cession

-0.65

agan

-0.62

 Yamato

-0.62

potion

-0.62

ading

-0.61

 liberated

-0.61

NT

-0.60

POSITIVE LOGITS

orial

0.89

heid

0.73

hyde

0.72

ledge

0.71

phrine

0.71

photo

0.68

auga

0.68

viron

0.67

emouth

0.66

ength

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.