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

aceous

-0.80

uman

-0.77

roma

-0.74

ideshow

-0.73

kefeller

-0.73

agu

-0.72

ulu

-0.71

culus

-0.71

ankind

-0.71

plets

-0.70

POSITIVE LOGITS

array

0.71

WR

0.65

 unfair

0.62

 Notice

0.62

Notice

0.61

ND

0.59

 debit

0.59

 missed

0.58

 Shrine

0.58

 Penalty

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.