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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

dylib

-0.72

metics

-0.70

 Sabha

-0.67

swick

-0.66

yna

-0.66

ngth

-0.66

gyn

-0.66

hus

-0.64

actionDate

-0.63

 ancest

-0.62

POSITIVE LOGITS

ÄŁ

0.67

ãĥķãĤ©

0.67

isu

0.66

Sol

0.65

 Portable

0.63

ãĥĩãĤ£

0.63

Zero

0.61

 Proced

0.61

Commercial

0.61

 Zero

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.