INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

antically

-0.66

bda

-0.65

irit

-0.63

butt

-0.61

Compat

-0.60

 cake

-0.60

ench

-0.60

ital

-0.60

irez

-0.59

ansk

-0.59

POSITIVE LOGITS

0.76

nut

0.71

los

0.71

ãĤ©

0.70

ãĥı

0.70

ãĥ¡

0.66

bleacher

0.66

urnal

0.63

Ĭ±

0.62

 loudspe

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.