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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

issance

-0.81

ospels

-0.78

edged

-0.75

isites

-0.74

soDeliveryDate

-0.72

AND

-0.69

beit

-0.68

leave

-0.67

antine

-0.66

imentary

-0.66

POSITIVE LOGITS

 assist

0.80

 helper

0.72

 helpers

0.70

 peasant

0.66

Attempts

0.63

 earners

0.61

 tutor

0.60

 peasants

0.60

 stands

0.59

å§«

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.