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

$$$$

-0.78

 mang

-0.68

isol

-0.66

 Kimber

-0.65

umph

-0.63

ishly

-0.62

Judge

-0.61

dragon

-0.61

interstitial

-0.60

 Nass

-0.59

POSITIVE LOGITS

endi

0.76

abi

0.70

hyde

0.68

abis

0.67

ooth

0.67

 âĸº

0.65

 Â®

0.65

velt

0.65

 gener

0.62

avia

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.