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

Animation

-0.77

ãĤ´ãĥ³

-0.70

ãĤ¨ãĥ«

-0.68

anamo

-0.66

TABLE

-0.66

 Rober

-0.65

 guid

-0.65

lihood

-0.62

arters

-0.62

yip

-0.62

POSITIVE LOGITS

gaard

0.74

stellar

0.71

ensical

0.68

dor

0.64

 pear

0.64

lite

0.64

avour

0.63

sylv

0.62

belt

0.62

 Shroud

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.