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

 Sailor

-0.65

BN

-0.65

RAY

-0.63

TPS

-0.63

ANS

-0.63

AIR

-0.62

EGIN

-0.62

CBS

-0.62

HAM

-0.62

BM

-0.62

POSITIVE LOGITS

 Practices

0.81

artifacts

0.67

enges

0.65

 Decay

0.65

 Causes

0.65

dll

0.64

 crus

0.64

 Machines

0.62

 pedals

0.62

edu

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.