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

 patched

-0.71

ModLoader

-0.63

 Vaugh

-0.63

âĶĢ

-0.63

isSpecialOrderable

-0.62

Davis

-0.62

 Conway

-0.62

 Appalach

-0.62

 Bolton

-0.61

condition

-0.60

POSITIVE LOGITS

ibal

0.82

anooga

0.75

iac

0.75

oint

0.71

 outp

0.70

prus

0.67

 Mechdragon

0.66

ramids

0.64

olith

0.64

pher

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.