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

rition

-0.74

entin

-0.68

ights

-0.67

umen

-0.65

asures

-0.65

ional

-0.65

igmat

-0.64

ritional

-0.64

owered

-0.63

 paraly

-0.63

POSITIVE LOGITS

 Chronicle

0.74

assetsadobe

0.73

geist

0.72

ĺħ

0.72

 Saud

0.69

vell

0.66

 Gaia

0.63

fell

0.61

 Learns

0.61

alde

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.