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

pless

-0.73

 mummy

-0.69

 Mechdragon

-0.68

incent

-0.67

IENT

-0.64

antly

-0.64

roud

-0.64

oslav

-0.63

 tarn

-0.61

toe

-0.61

POSITIVE LOGITS

Hack

0.76

Console

0.73

Anonymous

0.71

Writ

0.70

Statement

0.70

 Lists

0.70

ortium

0.69

Rules

0.68

ync

0.67

 Lisp

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.