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

lance

-0.74

uthor

-0.70

 lockdown

-0.69

oldown

-0.68

onite

-0.67

 proliferation

-0.65

 spectrum

-0.65

 presidency

-0.65

 Alphabet

-0.61

 grandson

-0.60

POSITIVE LOGITS

ãĤĬ

0.79

POST

0.71

æ©

0.71

buff

0.70

real

0.68

ãĥ¼ãĤ¯

0.68

ãĥŁ

0.68

)=(

0.67

painted

0.67

psc

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.