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

REDACTED

-0.70

interrupted

-0.70

hesive

-0.67

itionally

-0.67

ITED

-0.64

ohydrate

-0.63

 Gadget

-0.63

ebted

-0.62

SPONSORED

-0.62

Chel

-0.61

POSITIVE LOGITS

eco

0.70

pport

0.69

ĨĴ

0.66

NP

0.66

orient

0.65

IRC

0.64

umar

0.62

cases

0.61

 amnesty

0.60

 ener

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.