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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Sax

-0.73

 Oaks

-0.72

empt

-0.72

 Hastings

-0.68

eries

-0.68

wagon

-0.66

 Swed

-0.65

aston

-0.65

 Tend

-0.64

acre

-0.63

POSITIVE LOGITS

ordan

0.77

aturday

0.72

itars

0.68

ivered

0.65

overed

0.65

 escaping

0.64

abwe

0.62

soDeliveryDate

0.62

iliar

0.62

ascus

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.