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

TBD

-0.80

Ryan

-0.71

dropping

-0.71

continue

-0.68

hips

-0.67

ingly

-0.66

athan

-0.64

Alert

-0.64

posts

-0.63

JR

-0.62

POSITIVE LOGITS

¬¼

0.81

 Sorceress

0.79

 elder

0.76

 same

0.72

oples

0.70

sembly

0.69

 municipality

0.68

yss

0.68

 initiation

0.67

 aforementioned

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.