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

ĸļ

-0.90

interstitial

-0.77

URES

-0.74

sbm

-0.73

 EDITION

-0.71

DERR

-0.70

platform

-0.70

 Aires

-0.69

AUT

-0.69

LOC

-0.68

POSITIVE LOGITS

 perpet

0.76

 honor

0.69

vir

0.69

 gull

0.68

iety

0.66

 dishon

0.66

 dove

0.65

vantage

0.64

xes

0.63

ority

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.