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

Redditor

-0.78

POST

-0.76

cients

-0.73

ãĤ¦ãĤ¹

-0.71

DERR

-0.71

codes

-0.69

ACTION

-0.67

TEXTURE

-0.67

ICLE

-0.66

Cub

-0.66

POSITIVE LOGITS

 Albania

0.67

amy

0.63

Prime

0.62

 prime

0.61

 bankrupt

0.61

 bracket

0.61

 Fritz

0.60

 sabot

0.59

 Paula

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.