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

Merit

-0.79

Radio

-0.78

 Hutch

-0.77

NZ

-0.77

Boss

-0.77

Ctrl

-0.76

 Flavoring

-0.75

Room

-0.75

APD

-0.74

 Agents

-0.74

POSITIVE LOGITS

enne

0.79

bilt

0.76

[&

0.72

 magically

0.71

 shred

0.67

 hiber

0.66

dfx

0.65

 cocoa

0.65

 sperm

0.64

 semen

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.