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

natureconservancy

-0.89

 ILCS

-0.64

INTON

-0.62

 veto

-0.62

è¡

-0.61

 bowl

-0.61

NetMessage

-0.60

ãĥ¢

-0.60

 behaviours

-0.60

 ãĤµ

-0.60

POSITIVE LOGITS

©¶æ

0.86

witz

0.77

stadt

0.68

etsk

0.65

hler

0.65

asta

0.64

kus

0.63

allo

0.62

veyard

0.62

chnology

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.