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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

MpServer

-0.76

.")

-0.75

IAL

-0.71

acters

-0.70

externalActionCode

-0.68

 dancers

-0.66

mble

-0.64

ikawa

-0.62

 cocoa

-0.62

ICLE

-0.61

POSITIVE LOGITS

edia

0.77

ãĤ¨ãĥ«

0.70

 Dirt

0.67

 Seek

0.67

 Battlefield

0.67

 Disorder

0.67

Bound

0.66

 Vessel

0.64

æµ

0.64

oard

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.