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

 stalk

-0.78

 evapor

-0.70

 span

-0.62

 brim

-0.58

 Giul

-0.58

liest

-0.58

 burned

-0.58

 cooked

-0.57

Zan

-0.57

 brow

-0.56

POSITIVE LOGITS

pb

0.91

issance

0.75

backer

0.71

 Berm

0.71

itivity

0.70

ubi

0.70

uph

0.70

PLA

0.69

 Cosponsors

0.69

udi

0.69

Activations Density 0.000%

No Known Activations

This feature has no known activations.