INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

oo

-0.78

ynski

-0.76

ihad

-0.75

igen

-0.74

ception

-0.72

esome

-0.70

anson

-0.69

 geek

-0.67

zilla

-0.66

uction

-0.66

POSITIVE LOGITS

 Corruption

0.86

 Investigator

0.74

 Pipeline

0.74

psc

0.74

 Guards

0.72

 Shutdown

0.71

 Papers

0.71

 Investigative

0.70

 Regulations

0.69

endar

0.69

Activations Density 0.000%

No Known Activations

This feature has no known activations.