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

rouse

-0.66

natureconservancy

-0.62

});

-0.60

 prett

-0.60

 Â®

-0.59

 simplest

-0.59

rul

-0.59

ills

-0.58

]=

-0.58

 foreseeable

-0.58

POSITIVE LOGITS

irlf

0.81

omething

0.71

asuring

0.70

alian

0.69

 about

0.68

raltar

0.68

interface

0.67

 Magnet

0.66

oslov

0.65

byss

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.