INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

GES

-0.76

FC

-0.71

 ç¥ŀ

-0.71

 Grade

-0.70

ennes

-0.70

SE

-0.69

TOP

-0.68

NE

-0.66

Favorite

-0.66

Stock

-0.63

POSITIVE LOGITS

azeera

0.74

wikipedia

0.72

urnal

0.71

 volunt

0.67

itialized

0.66

pointer

0.65

 Authors

0.65

pora

0.64

 maze

0.63

 analog

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.