INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

 arrang

-0.80

pring

-0.75

 veter

-0.69

åĤ

-0.69

 tabl

-0.69

liga

-0.64

 enthusi

-0.64

 Parables

-0.62

 imitate

-0.61

 accur

-0.61

POSITIVE LOGITS

ihad

0.77

ibus

0.77

iden

0.74

 GOODMAN

0.71

UCT

0.69

Ground

0.68

elson

0.67

 Campus

0.67

IVERS

0.66

OST

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.