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

 consultants

-0.75

adesh

-0.70

 Merlin

-0.68

 buckle

-0.68

 Courier

-0.67

 whis

-0.65

ecast

-0.65

GPS

-0.63

fax

-0.63

 retract

-0.63

POSITIVE LOGITS

nces

0.68

Fall

0.65

ANCE

0.64

UCT

0.64

Guest

0.62

gradation

0.62

achel

0.61

uctions

0.61

Viol

0.61

ANC

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.