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

etsk

-0.83

inki

-0.78

zh

-0.78

Rus

-0.76

ovo

-0.76

sided

-0.76

edin

-0.72

igrated

-0.69

igrate

-0.69

 Frie

-0.67

POSITIVE LOGITS

ģĸ

0.69

 plumbing

0.64

Ĥª

0.63

 Clojure

0.61

ubes

0.61

 commod

0.61

 thesis

0.60

 Expend

0.60

 framing

0.59

 exponent

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.