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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Cir

-0.68

 Soros

-0.67

 Burr

-0.65

Rai

-0.65

 Riding

-0.63

 Slot

-0.62

Ort

-0.60

 Irwin

-0.59

 Forbes

-0.59

 Passive

-0.58

POSITIVE LOGITS

ysis

0.86

illin

0.86

enh

0.85

isha

0.75

eeper

0.71

peror

0.70

ibo

0.70

fax

0.70

itous

0.69

obyl

0.68

Activations Density 0.000%

No Known Activations

This feature has no known activations.