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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

oreal

-0.83

VIDIA

-0.79

ategory

-0.79

ivating

-0.77

ruary

-0.77

uay

-0.76

 Palest

-0.75

arling

-0.75

ivated

-0.73

ivariate

-0.73

POSITIVE LOGITS

bye

0.75

 Isles

0.67

 Finder

0.67

Tid

0.66

DNA

0.61

 Promise

0.61

 Franks

0.60

 resp

0.60

é¾į

0.60

 Names

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.