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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

defense

-0.64

-0.59

-0.58

 âĢİ

-0.58

lun

-0.58

 Bridgewater

-0.57

 Barkley

-0.57

 replied

-0.57

mental

-0.56

mag

-0.56

POSITIVE LOGITS

Ô

1.05

mble

0.87

Cth

0.87

 è£ıè

0.84

ħĭ

0.79

ADRA

0.75

ãĤ´

0.74

iple

0.74

stasy

0.73

culosis

0.72

Activations Density 0.000%

No Known Activations

This feature has no known activations.