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

 Strength

-0.73

 ultras

-0.70

 abundance

-0.68

"+

-0.65

fir

-0.63

 clarity

-0.62

 centre

-0.62

-|

-0.61

Firstly

-0.60

del

-0.60

POSITIVE LOGITS

merce

0.81

mercial

0.78

ModLoader

0.78

legram

0.77

occ

0.75

eday

0.75

wald

0.72

olitics

0.69

veyard

0.68

ICLE

0.68

Activations Density 0.000%

No Known Activations

This feature has no known activations.