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

HOU

-0.80

XXX

-0.76

heimer

-0.74

ufact

-0.71

iple

-0.69

aurus

-0.68

ARI

-0.68

teenth

-0.67

HB

-0.66

arte

-0.66

POSITIVE LOGITS

llular

0.83

*/(

0.72

 takeaway

0.70

 antim

0.68

ndra

0.67

 creatine

0.65

 tweet

0.65

 hars

0.65

 longer

0.65

agnetic

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.