INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

enegger

-0.77

iago

-0.73

Enough

-0.72

acters

-0.67

aukee

-0.67

ween

-0.66

 Azerb

-0.66

 Kers

-0.65

atar

-0.64

eri

-0.64

POSITIVE LOGITS

isSpecialOrderable

0.72

 Pilgrim

0.71

itent

0.70

 instincts

0.68

 bonded

0.64

ifles

0.64

natureconservancy

0.64

acious

0.63

rouse

0.61

 bars

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.