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

painted

-0.66

 begg

-0.63

yes

-0.61

bole

-0.61

feet

-0.59

Ã°

-0.58

PLIED

-0.58

gel

-0.58

 Kazakh

-0.57

zik

-0.57

POSITIVE LOGITS

ombat

0.88

FTWARE

0.80

-+-+

0.77

uliffe

0.74

 sequels

0.72

ombies

0.71

bernatorial

0.70

glomer

0.70

LR

0.69

 +---

0.68

Activations Density 0.000%

No Known Activations

This feature has no known activations.