INDEX

Explanations

references to promotional materials or advertising elements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 29-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L29R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.29.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

sWith

-0.39

sel

-0.38

sm

-0.36

sin

-0.35

side

-0.35

sp

-0.35

sc

-0.34

sis

-0.34

sid

-0.34

sen

-0.33

POSITIVE LOGITS

idge

0.35

er

0.31

ë§ģ

0.30

ific

0.29

gebn

0.29

ized

0.28

cury

0.28

ed

0.28

lain

0.27

most

0.26

Activations Density 0.732%

references to promotional materials or advertising elements

No Comments

No Known Activations