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

Ó

-0.72

icked

-0.71

imes

-0.71

gets

-0.71

ventus

-0.71

ewitness

-0.69

hess

-0.68

soDeliveryDate

-0.68

ahoo

-0.68

rax

-0.67

POSITIVE LOGITS

erton

0.67

TPP

0.66

 Chrome

0.64

 upward

0.63

MIT

0.62

 NAFTA

0.61

 plumbing

0.60

 extradition

0.60

 FreeBSD

0.60

 parenting

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.