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

via

-0.79

hur

-0.70

jah

-0.70

regon

-0.68

meet

-0.67

mph

-0.67

icy

-0.67

:]

-0.66

hip

-0.65

cum

-0.65

POSITIVE LOGITS

 Chronicles

0.73

obos

0.64

 Stur

0.64

ierre

0.63

 bindings

0.63

 insign

0.62

 Amend

0.61

 Conquest

0.61

 recoil

0.61

 McGu

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.