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

ISM

-0.74

 Helpful

-0.74

 Squid

-0.72

 Blizz

-0.70

 Haram

-0.69

 Reader

-0.65

ictions

-0.65

 Catalonia

-0.64

 Trader

-0.63

Hik

-0.62

POSITIVE LOGITS

ramid

0.70

modules

0.70

oster

0.67

itially

0.65

PRO

0.63

 Rece

0.63

ixel

0.61

 skelet

0.60

eria

0.60

kees

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.