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

 knees

-0.71

 exempted

-0.67

 computed

-0.65

doms

-0.64

utherland

-0.63

 authored

-0.62

 earned

-0.61

 indexed

-0.61

zed

-0.60

thritis

-0.60

POSITIVE LOGITS

rio

0.71

eps

0.71

 Olympia

0.68

è¦ļéĨĴ

0.67

↵

0.66

bike

0.66

abi

0.65

 Dream

0.65

ËĪ

0.64

 Thousand

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.