INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ongyang

-0.78

onga

-0.75

imaru

-0.74

hesda

-0.74

ksh

-0.73

bidden

-0.70

rants

-0.68

lest

-0.67

teen

-0.67

anooga

-0.67

POSITIVE LOGITS

ucer

0.72

 Pixie

0.70

 refuel

0.68

 mistaken

0.66

 externalToEVAOnly

0.66

 Toll

0.64

 Kard

0.63

ļéĨĴ

0.61

 Samoa

0.60

 Sorce

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.