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

ailability

-0.73

antes

-0.70

 Rolls

-0.69

Quotes

-0.68

RPG

-0.68

table

-0.67

 riches

-0.66

ript

-0.65

cript

-0.65

Newsletter

-0.64

POSITIVE LOGITS

 msec

0.75

ichen

0.70

 Whoever

0.69

>[

0.66

owicz

0.65

 princ

0.64

oub

0.62

 shutter

0.61

icter

0.61

icka

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.