INDEX

Explanations

references to specific entities or subjects, predominantly in a German context

oai_token-act-pair · gpt-4o-mini Triggered by @bot

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

igkeit

-0.17

taire

-0.16

AccessException

-0.15

izer

-0.15

Â§Ãĥ

-0.15

arend

-0.14

yr

-0.14

yre

-0.14

sd

-0.14

è¾

-0.14

POSITIVE LOGITS

anged

0.28

angement

0.24

ivative

0.23

ksen

0.22

went

0.21

ivation

0.20

neÄŁi

0.19

udder

0.19

ivery

0.19

wish

0.18

Activations Density 0.014%

references to specific entities or subjects, predominantly in a German context

No Comments

No Known Activations