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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

trop

-0.84

icum

-0.75

 Blizz

-0.69

 Strongh

-0.68

 Enhancement

-0.68

haar

-0.67

âĹ¼

-0.67

tnc

-0.67

afety

-0.67

acea

-0.65

POSITIVE LOGITS

INGTON

0.65

ÑĮ

0.64

æ©Ł

0.63

äºĶ

0.61

ishi

0.60

LM

0.60

ãĥı

0.60

Ħ

0.58

 à¤

0.57

itialized

0.57

Activations Density 0.000%

No Known Activations

This feature has no known activations.