INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

sey

-0.76

say

-0.70

bare

-0.69

athing

-0.68

agh

-0.68

lining

-0.66

ving

-0.66

utsche

-0.66

ryu

-0.65

liner

-0.65

POSITIVE LOGITS

ï¸

0.85

é¾įåĸļå£«

0.75

natureconservancy

0.67

NetMessage

0.66

Effective

0.65

Services

0.65

âĶĢâĶĢâĶĢâĶĢâĶĢâĶĢâĶĢâĶĢ

0.63

 cran

0.63

Ped

0.62

 externalToEVAOnly

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.