INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

nai

-0.90

ktop

-0.87

aido

-0.68

 charism

-0.66

Leaks

-0.65

iment

-0.64

ivo

-0.64

nuts

-0.64

assy

-0.63

 electr

-0.62

POSITIVE LOGITS

othal

0.70

 transmissions

0.65

 Revenge

0.64

 Trojan

0.63

 treasure

0.61

 scrimmage

0.60

 salute

0.60

oS

0.59

=-=-=-=-

0.59

 GEAR

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.