INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 30-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L30R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.30.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

cerebras/SlimPajama-627B

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Kay

-0.16

idor

-0.15

nete

-0.15

 lyon

-0.15

 Franklin

-0.14

otta

-0.14

opa

-0.14

GBT

-0.14

Kay

-0.13

POSITIVE LOGITS

 carbs

0.23

 ketogenic

0.22

 carb

0.21

aqu

0.21

 diet

0.21

 Carb

0.21

 landslide

0.20

 Diet

0.20

 carbohydrate

0.19

 groundwater

0.19

Activations Density 0.000%

No Known Activations

This feature has no known activations.