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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

efully

-0.72

abilia

-0.69

ingly

-0.67

iannopoulos

-0.66

owship

-0.66

jamin

-0.65

jri

-0.65

ILY

-0.65

colored

-0.63

 Ellen

-0.62

POSITIVE LOGITS

maxwell

0.77

vine

0.76

ury

0.67

 Fever

0.65

arers

0.62

 clipboard

0.61

 Fury

0.60

Byr

0.60

 heartbeat

0.60

 bottleneck

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.