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

 ï¿½

-0.75

hari

-0.74

lish

-0.69

ules

-0.67

 Flavoring

-0.67

 ILCS

-0.67

OTOS

-0.66

LOD

-0.65

ulf

-0.64

ional

-0.64

POSITIVE LOGITS

beit

0.75

shop

0.74

vantage

0.70

choes

0.65

 voic

0.62

pi

0.61

 strikers

0.60

 seiz

0.60

izon

0.58

 frustrated

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.