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

retty

-0.77

ppel

-0.69

gex

-0.68

Fight

-0.68

acters

-0.64

ourke

-0.63

draft

-0.63

pb

-0.63

akedown

-0.60

ymes

-0.60

POSITIVE LOGITS

 Flavoring

0.74

endi

0.71

Wan

0.67

atron

0.65

Ø§ÙĦ

0.64

>>>>>>>>

0.62

abeth

0.61

ety

0.61

TPPStreamerBot

0.61

cu

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.