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

-0.82

osponsors

-0.79

ersed

-0.72

ascript

-0.72

bringer

-0.71

haul

-0.71

ragon

-0.71

Pwr

-0.68

arthed

-0.68

ilogy

-0.67

POSITIVE LOGITS

 Head

0.68

 mouths

0.65

 Province

0.62

itar

0.62

EV

0.62

 Environment

0.62

IO

0.61

 Manager

0.60

 pipes

0.60

Gas

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.