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

æĪ¦

-0.70

ilde

-0.68

Translation

-0.67

 Nanto

-0.66

 Dresden

-0.65

 theolog

-0.65

 lieutenant

-0.65

 Bundes

-0.64

 nominal

-0.64

åĽ

-0.63

POSITIVE LOGITS

 where

1.19

 whence

1.04

where

0.87

thens

0.71

icularly

0.70

ents

0.70

vati

0.69

 wherein

0.69

fman

0.66

 WHERE

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.