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

anooga

-0.71

awaru

-0.68

anmar

-0.67

ratulations

-0.63

 gamb

-0.62

Ò

-0.62

ja

-0.61

aido

-0.60

 jerk

-0.59

amiya

-0.58

POSITIVE LOGITS

urn

0.78

WATCH

0.75

TED

0.72

thumbnails

0.72

READ

0.72

Interstitial

0.69

ITION

0.68

 Lauder

0.68

~/.

0.67

ined

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.