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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 adultery

-0.79

ħĭ

-0.69

 probation

-0.69

arius

-0.68

 detract

-0.65

hess

-0.64

Õ

-0.63

atis

-0.62

 persecut

-0.62

 infring

-0.61

POSITIVE LOGITS

enhagen

0.77

 SERV

0.69

ocks

0.67

 Location

0.65

>]

0.65

ï¸ı

0.64

 Posted

0.63

 transports

0.63

roleum

0.63

cham

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.