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

entric

-0.75

OLOG

-0.72

ological

-0.65

ove

-0.65

gling

-0.65

HC

-0.63

 Benz

-0.62

oxy

-0.61

ulate

-0.60

,—

-0.59

POSITIVE LOGITS

taboola

0.86

Brexit

0.80

 notor

0.79

markets

0.79

©¶æ¥µ

0.78

 trave

0.71

abouts

0.69

 exting

0.68

Europe

0.67

ndra

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.