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

ãĤ¨ãĥ«

-0.68

apo

-0.67

Language

-0.66

 Athe

-0.63

apy

-0.63

polit

-0.61

Route

-0.59

istani

-0.58

LL

-0.58

ready

-0.57

POSITIVE LOGITS

[|

0.68

 comr

0.67

ĪĴ

0.64

 refunds

0.63

 Drum

0.63

xual

0.62

semble

0.62

 discounts

0.62

igans

0.62

interstitial

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.