INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 Gors

-0.76

manship

-0.70

company

-0.66

stal

-0.64

hold

-0.63

 playbook

-0.61

CLS

-0.61

young

-0.60

SK

-0.60

Emb

-0.60

POSITIVE LOGITS

berra

0.88

igi

0.80

redits

0.77

foreseen

0.75

aturdays

0.67

hran

0.65

agnar

0.65

reenshots

0.63

vae

0.63

asio

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.