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

wagen

-0.77

rez

-0.74

Chel

-0.73

-0.70

IPM

-0.67

di

-0.66

chlor

-0.66

raq

-0.65

vable

-0.64

lehem

-0.63

POSITIVE LOGITS

ativity

0.67

 Dickinson

0.64

 Alger

0.61

 proposition

0.61

 fright

0.60

 hemp

0.60

 peaks

0.60

hammer

0.59

 Perry

0.59

eah

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.