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

çīĪ

-0.74

devices

-0.68

athed

-0.62

 Corinth

-0.61

paralle

-0.61

 unequ

-0.60

 Bezos

-0.60

ģ«

-0.58

Hew

-0.58

 Indra

-0.58

POSITIVE LOGITS

vision

0.69

rador

0.68

umblr

0.64

esy

0.64

IUM

0.64

querque

0.63

 SCHOOL

0.63

ilial

0.63

ONY

0.61

edin

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.