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

onga

-0.92

zel

-0.75

ickets

-0.75

emort

-0.70

ographically

-0.68

zig

-0.67

atro

-0.67

ingen

-0.67

enum

-0.66

awar

-0.65

POSITIVE LOGITS

CTV

0.81

af

0.70

 Invalid

0.67

 pony

0.66

ascade

0.65

 Moff

0.64

 anew

0.64

fm

0.64

CET

0.64

coni

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.