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

etheless

-0.71

izabeth

-0.71

ample

-0.71

¶ħ

-0.69

wagen

-0.69

alky

-0.67

rette

-0.65

examination

-0.64

ueller

-0.63

creen

-0.63

POSITIVE LOGITS

ership

0.66

erness

0.63

 Samar

0.63

 Shots

0.60

0.59

 Cargo

0.59

 Plains

0.59

 Alive

0.58

 Radical

0.58

Anim

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.