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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

obal

-0.78

µ

-0.74

 destro

-0.69

ciating

-0.68

uana

-0.68

ivot

-0.68

zb

-0.66

monds

-0.64

±

-0.63

ython

-0.62

POSITIVE LOGITS

gyn

0.64

 Fritz

0.62

 Strauss

0.61

fred

0.60

achus

0.59

heimer

0.59

vier

0.58

born

0.58

 Babe

0.58

 Conservative

0.57

Activations Density 0.000%

No Known Activations

This feature has no known activations.