INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 Norn

-0.75

 mats

-0.73

åĨ

-0.69

asse

-0.69

company

-0.68

"},"

-0.67

hoe

-0.66

MGM

-0.65

/>

-0.64

atre

-0.63

POSITIVE LOGITS

assium

0.71

 smuggling

0.70

ktop

0.69

abulary

0.65

chedel

0.64

senal

0.64

 experimentation

0.63

earances

0.62

ruciating

0.62

 deserved

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.