INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

rity

-1.06

opes

-0.85

cair

-0.80

ippi

-0.78

udence

-0.72

uters

-0.71

omon

-0.71

ocaly

-0.70

 guiActiveUn

-0.69

utsche

-0.69

POSITIVE LOGITS

 scrimmage

0.72

Adult

0.72

 sund

0.67

Redditor

0.66

 hereafter

0.62

whe

0.59

Untitled

0.59

within

0.59

 ration

0.58

Ve

0.57

Activations Density 0.000%

No Known Activations

This feature has no known activations.