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

CTV

-0.82

ÃŁ

-0.79

Ô

-0.78

thread

-0.74

atur

-0.71

okane

-0.70

iland

-0.70

utsch

-0.69

Veter

-0.69

%]

-0.69

POSITIVE LOGITS

 globe

0.74

 targ

0.73

 prone

0.71

 suicidal

0.67

 world

0.67

 murd

0.63

geries

0.62

 contrace

0.62

 suspic

0.62

 undermin

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.