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

 Grimm

-0.76

 Rangers

-0.66

 Quran

-0.65

 scapego

-0.62

 cant

-0.62

 wors

-0.58

 Lotus

-0.58

 autom

-0.57

 demons

-0.57

 Sent

-0.56

POSITIVE LOGITS

wcs

0.77

ahime

0.74

itle

0.70

 guiActiveUnfocused

0.70

Present

0.70

WARD

0.67

axter

0.67

Ã¯ve

0.65

cision

0.65

EStreamFrame

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.