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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ater

-0.73

 annex

-0.72

¾

-0.72

è¦ļéĨĴ

-0.71

DragonMagazine

-0.69

ãĥĥãĥĪ

-0.67

ieri

-0.66

ãĤ´ãĥ³

-0.65

HCR

-0.65

rawdownloadcloneembedreportprint

-0.65

POSITIVE LOGITS

edient

0.75

Bom

0.74

 Jenn

0.72

uesday

0.70

alach

0.69

 Blink

0.68

 Pike

0.67

 Comedy

0.66

ousy

0.65

 Ender

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.