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

ãĤ´

-0.83

REDACTED

-0.81

Cop

-0.74

Charl

-0.71

å¼

-0.70

Merc

-0.69

ãĤ¹

-0.69

rest

-0.67

äº

-0.67

äºĶ

-0.66

POSITIVE LOGITS

enario

0.78

anza

0.72

ython

0.72

imates

0.71

iary

0.68

olate

0.67

}}}

0.67

arios

0.66

zees

0.66

 hike

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.