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

 Hancock

-0.76

wich

-0.71

bah

-0.66

aney

-0.64

 Bunker

-0.62

enium

-0.62

chel

-0.60

buster

-0.60

 Straw

-0.60

ATI

-0.59

POSITIVE LOGITS

rists

0.75

ntil

0.71

LCS

0.67

ĺħ

0.67

 goalt

0.67

REF

0.67

FORE

0.64

ailability

0.62

ittens

0.62

 goaltender

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.