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

 Manit

-0.77

 pane

-0.70

oux

-0.69

 Waterloo

-0.64

Alz

-0.64

 mushrooms

-0.62

meric

-0.62

Eth

-0.62

conom

-0.62

POSITIVE LOGITS

HB

0.73

CVE

0.73

kefeller

0.69

itans

0.64

ravis

0.63

RL

0.63

ERAL

0.62

laus

0.62

 caric

0.61

DEF

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.