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

 Lerner

-0.74

 Bundy

-0.72

deals

-0.69

uten

-0.68

deal

-0.66

stal

-0.63

 Steele

-0.63

****************

-0.63

 subsidiary

-0.62

 Sheila

-0.62

POSITIVE LOGITS

ħĭ

0.71

 childbirth

0.70

Zed

0.69

erker

0.68

ijn

0.67

 fashionable

0.66

 sshd

0.66

utterstock

0.65

rame

0.65

itbart

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.