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

 WITHOUT

-0.68

iji

-0.68

 compression

-0.68

 clipboard

-0.67

ãĥĥãĥī

-0.65

iyah

-0.64

ãĥĥãĥĪ

-0.62

iety

-0.61

`,

-0.61

theless

-0.60

POSITIVE LOGITS

mas

0.71

anqu

0.68

atories

0.68

Fake

0.66

Fab

0.64

onomy

0.63

atisf

0.63

enda

0.60

 kilomet

0.59

 srfAttach

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.