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

ufact

-0.78

inous

-0.70

oros

-0.70

hesis

-0.69

 decre

-0.69

 circ

-0.67

mur

-0.65

unicip

-0.64

yss

-0.64

endas

-0.63

POSITIVE LOGITS

Reading

0.73

 latitude

0.71

ħĭ

0.70

Auth

0.66

ISA

0.66

 Background

0.63

Appearances

0.62

 Behavior

0.62

Testing

0.61

RH

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.