INDEX

Explanations

references to military-related themes and terminology

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 31-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L31R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.31.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

ãĤ¡

-0.14

ajar

-0.14

èĦĳ

-0.13

Î³Ïī

-0.13

 Entire

-0.13

Misc

-0.13

eza

-0.13

 ë§ĲìĿ´

-0.13

yar

-0.12

[__

-0.12

POSITIVE LOGITS

 ranks

0.35

 fold

0.34

 workforce

0.30

 fray

0.28

 employ

0.27

fold

0.26

 Fold

0.24

 folds

0.24

 priesthood

0.24

 trenches

0.23

Activations Density 0.144%

references to military-related themes and terminology

No Comments

No Known Activations