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

ILLE

-0.84

utsche

-0.74

ersive

-0.71

acion

-0.68

ATIVE

-0.68

uder

-0.67

orescent

-0.67

Mobile

-0.66

USB

-0.65

isner

-0.64

POSITIVE LOGITS

etheless

0.69

 inclination

0.67

terday

0.66

awareness

0.65

 intention

0.64

[/

0.63

 drowning

0.62

 prone

0.62

yang

0.60

 carp

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.