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

 dwellings

-0.69

 flakes

-0.66

 Ital

-0.63

cules

-0.62

 houses

-0.61

 circles

-0.60

 digits

-0.59

terness

-0.59

 tons

-0.59

 periphery

-0.59

POSITIVE LOGITS

iHUD

0.86

ellen

0.85

 misunder

0.79

\\\\\\\\

0.74

 warr

0.71

âĶĢâĶĢâĶĢâĶĢ

0.71

////////////////

0.67

Afee

0.66

 Notting

0.65

 Ples

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.