INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

tight

-0.63

abal

-0.62

breakers

-0.60

kees

-0.60

interstitial

-0.60

 filler

-0.60

zeb

-0.59

...]

-0.59

wall

-0.58

 embr

-0.58

POSITIVE LOGITS

idium

0.70

****

0.65

Downloadha

0.64

etooth

0.63

edom

0.62

itudes

0.62

GoldMagikarp

0.61

employment

0.61

 ACTION

0.61

*/(

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.