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

POS

-0.68

CTR

-0.65

finder

-0.63

FN

-0.62

proof

-0.62

annie

-0.62

ournal

-0.62

lod

-0.60

 EntityItem

-0.60

Guide

-0.59

POSITIVE LOGITS

vol

1.61

Sep

0.90

vo

0.72

vy

0.71

 Volume

0.64

azel

0.64

ashi

0.63

ime

0.63

vag

0.61

ahl

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.