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

 Cooperation

-0.69

interstitial

-0.67

Eid

-0.67

 Collective

-0.66

naire

-0.64

 Tycoon

-0.62

kin

-0.62

 cooperative

-0.61

 Import

-0.61

bite

-0.61

POSITIVE LOGITS

ãĤ¼ãĤ¦ãĤ¹

0.78

iannopoulos

0.72

essional

0.68

iatus

0.67

 COUR

0.66

itably

0.66

inis

0.65

 fuckin

0.65

ERC

0.64

mast

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.