INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

 uncond

-0.70

xon

-0.69

 Doll

-0.63

phrine

-0.62

etsk

-0.62

 Means

-0.62

 Terr

-0.61

 manifests

-0.60

GUN

-0.59

-------

-0.59

POSITIVE LOGITS

Dub

0.72

è£ħ

0.71

amy

0.70

aida

0.70

Advertisements

0.69

 Slam

0.67

 Chic

0.67

å§

0.65

imoto

0.65

Setup

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.