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

sett

-0.86

 behalf

-0.72

hement

-0.65

PLIED

-0.65

ryn

-0.64

kb

-0.64

rongh

-0.64

umar

-0.63

Recommend

-0.62

ulus

-0.62

POSITIVE LOGITS

 fixtures

0.73

 âĶľ

0.70

 BART

0.63

 gears

0.62

 queens

0.61

 Rodrigo

0.61

 bunny

0.60

 ladder

0.59

oÄŁ

0.59

 Races

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.