INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 26-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L26R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.26.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

-0.15

-0.14

 Episcopal

-0.14

èĩ

-0.14

-0.13

illin

-0.13

ÙĳØ©

-0.13

POSITIVE LOGITS

jam

0.24

Jam

0.21

pad

0.17

jam

0.17

jaz

0.16

Jam

0.16

 Mush

0.16

Pad

0.16

 jams

0.15

Activations Density 0.000%

No Known Activations

This feature has no known activations.