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

shield

-0.71

olor

-0.69

ricks

-0.64

reed

-0.61

 Yard

-0.60

cffffcc

-0.60

 stash

-0.60

FML

-0.59

 Watts

-0.58

lde

-0.58

POSITIVE LOGITS

 Apex

0.69

appropriately

0.67

 festive

0.65

çķ

0.64

 Ragnarok

0.64

 author

0.64

ache

0.63

ascript

0.62

 dign

0.62

natureconservancy

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.