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

 msec

-0.72

encers

-0.65

LTD

-0.65

 Noble

-0.64

 HOUSE

-0.63

zynski

-0.63

 Springs

-0.60

lev

-0.60

kW

-0.59

iya

-0.59

POSITIVE LOGITS

 Typh

0.74

 eyeb

0.69

 Canaver

0.69

Buff

0.69

Aren

0.68

atche

0.67

opard

0.64

geon

0.63

 Arbor

0.63

ahoo

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.