INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

ios

-0.85

 mathemat

-0.74

EMA

-0.72

Ã³n

-0.71

uthor

-0.69

oche

-0.69

ouse

-0.69

 mosqu

-0.68

asive

-0.67

ived

-0.66

POSITIVE LOGITS

Tip

0.86

 Nash

0.64

 Strikes

0.64

 Dahl

0.63

 libertarians

0.62

 Liber

0.61

 Foley

0.61

Dag

0.58

 Weir

0.58

 Bundy

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.