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

anza

-0.78

itiz

-0.74

esame

-0.72

ances

-0.71

fal

-0.69

aimon

-0.69

requently

-0.68

orns

-0.68

itely

-0.67

ancing

-0.66

POSITIVE LOGITS

 Bonnie

0.71

 Squid

0.70

 Lilly

0.70

 Adidas

0.67

 Slayer

0.65

enhagen

0.65

 Hitman

0.64

 Mutant

0.64

 RELEASE

0.63

 Speedway

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.