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

 Atlantis

-0.75

 Cosponsors

-0.73

CVE

-0.70

 Remastered

-0.68

ALD

-0.68

Lv

-0.65

ABLE

-0.65

chanted

-0.65

 Harmony

-0.65

 Revival

-0.61

POSITIVE LOGITS

erate

0.71

ortium

0.69

igroup

0.69

icides

0.68

estamp

0.67

aws

0.67

gres

0.65

eteen

0.65

sect

0.65

icol

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.