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

Pwr

-0.86

 acknow

-0.75

VID

-0.71

YC

-0.70

pex

-0.69

\-

-0.69

CVE

-0.68

TEXT

-0.68

STA

-0.68

udic

-0.67

POSITIVE LOGITS

ius

0.77

 Mith

0.68

iously

0.66

 republic

0.64

iang

0.64

 communism

0.61

 Rand

0.61

ocry

0.61

cation

0.60

 unrest

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.