INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

avers

-0.77

 suspic

-0.76

eria

-0.74

erker

-0.71

oresc

-0.70

uchi

-0.68

everal

-0.67

piring

-0.67

 succession

-0.67

ucker

-0.65

POSITIVE LOGITS

 Surf

0.73

 Annotations

0.67

 airs

0.67

 jails

0.65

 mirrors

0.64

jar

0.64

 Sadd

0.63

 Intercept

0.63

Sri

0.62

 Newsletter

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.