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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

cffff

-0.76

housing

-0.66

MpServer

-0.62

ModLoader

-0.62

Adds

-0.61

 Letters

-0.61

 Gears

-0.60

govtrack

-0.60

 Scouting

-0.60

è£ħ

-0.60

POSITIVE LOGITS

afe

0.83

è¦ļéĨĴ

0.72

oke

0.70

ierre

0.67

ahan

0.66

borg

0.65

omal

0.65

urus

0.65

phi

0.64

imate

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.