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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

âĸ¬

-0.66

urous

-0.63

ochet

-0.63

 throats

-0.62

atin

-0.61

ble

-0.61

apo

-0.59

ilet

-0.59

itton

-0.58

ocon

-0.58

POSITIVE LOGITS

DragonMagazine

0.94

riott

0.76

eous

0.69

Store

0.69

users

0.68

"}

0.67

soDeliveryDate

0.66

Reviewer

0.65

Sound

0.65

thodox

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.