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

natureconservancy

-0.81

Quotes

-0.79

TextColor

-0.73

clips

-0.67

Oracle

-0.65

 hell

-0.64

afort

-0.64

è£

-0.62

 sovere

-0.61

CVE

-0.61

POSITIVE LOGITS

endon

0.91

nder

0.86

ophile

0.77

ocial

0.72

ural

0.69

ement

0.66

ogie

0.66

Miy

0.66

tein

0.66

 Franchise

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.