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

enta

-0.88

essen

-0.81

liga

-0.77

ornia

-0.74

onis

-0.73

pillar

-0.68

rity

-0.67

aspers

-0.66

auri

-0.66

rite

-0.65

POSITIVE LOGITS

 Lost

0.65

kj

0.64

 unhappy

0.64

 Luffy

0.63

uctor

0.60

ODUCT

0.59

DK

0.58

 Token

0.57

 Osaka

0.56

 Hide

0.56

Activations Density 0.000%

No Known Activations

This feature has no known activations.