INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

hyde

-0.64

ank

-0.64

rolet

-0.64

 sarc

-0.61

 runner

-0.57

aceous

-0.56

 sequ

-0.56

 fanbase

-0.56

 derog

-0.56

trak

-0.56

POSITIVE LOGITS

©¶æ

0.86

Reviewed

0.76

 Agric

0.71

iculture

0.68

Ü

0.66

emis

0.64

arat

0.64

 Schwarz

0.63

nen

0.63

[+

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.