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

¥ŀ

-0.73

 intermittent

-0.69

 peripher

-0.68

 outstanding

-0.67

 intermitt

-0.67

 unbearable

-0.67

 contingency

-0.67

 unfinished

-0.66

 bottleneck

-0.65

 spor

-0.64

POSITIVE LOGITS

arta

0.86

article

0.84

afe

0.71

ukong

0.69

 Viet

0.69

ortex

0.69

letico

0.68

 ILCS

0.67

umbnail

0.67

ados

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.