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

ICLE

-0.84

APTER

-0.74

rawdownloadcloneembedreportprint

-0.73

Tube

-0.73

ilee

-0.70

 Baton

-0.70

EngineDebug

-0.70

EVA

-0.67

è£ħ

-0.67

afety

-0.67

POSITIVE LOGITS

ursed

0.77

 Blumenthal

0.70

 Ernst

0.69

 goblin

0.67

established

0.66

azi

0.65

Sax

0.65

xe

0.65

Ãº

0.65

 Rothschild

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.