INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

ICLE

-0.74

ICS

-0.71

 contestant

-0.68

wan

-0.66

ISM

-0.66

ISE

-0.64

ä¿

-0.64

çİĭ

-0.64

CP

-0.62

STER

-0.61

POSITIVE LOGITS

MpServer

0.78

 Slovenia

0.72

 Syri

0.71

iets

0.70

 Sloven

0.70

Jes

0.68

igraph

0.68

emonium

0.67

 headache

0.66

 teasp

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.