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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ilings

-0.82

 emanc

-0.80

 caric

-0.73

acebook

-0.70

 projection

-0.67

 indo

-0.67

 opin

-0.67

ODUCT

-0.67

 analytic

-0.66

 guiActiveUnfocused

-0.66

POSITIVE LOGITS

 Butcher

0.79

 McMahon

0.71

 Williamson

0.71

whe

0.69

 Ogre

0.69

mia

0.69

 Meyer

0.67

 Graves

0.67

Ball

0.67

agonist

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.