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

 Breach

-0.66

 ACTIONS

-0.64

othy

-0.63

OHN

-0.62

illary

-0.62

ittees

-0.60

atern

-0.60

 Garrison

-0.60

ustomed

-0.59

 Painter

-0.58

POSITIVE LOGITS

lehem

0.69

kill

0.68

roller

0.68

 achie

0.68

 bells

0.67

 laure

0.66

aires

0.66

 sweets

0.65

apest

0.65

acular

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.