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

ue

-0.68

DAC

-0.61

 Assassins

-0.61

 confinement

-0.60

ague

-0.60

 detain

-0.59

 entry

-0.59

 dealers

-0.59

Osc

-0.59

 Condition

-0.59

POSITIVE LOGITS

ãĥķãĤ©

0.85

ãĤ¢ãĥ«

0.84

è»

0.82

ãĥĩãĤ£

0.81

åĩ

0.80

æµ

0.80

lord

0.77

icist

0.77

å¥³

0.75

erity

0.75

Activations Density 0.000%

No Known Activations

This feature has no known activations.