INDEX

Explanations

events related to violence or assaults involving weapons

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

ummer

-0.16

enance

-0.15

ãģ¡ãģ¯

-0.15

tember

-0.14

 Brigade

-0.14

rides

-0.14

ç½ļ

-0.14

_TP

-0.14

isson

-0.14

 ÑģÐµÐ·

-0.13

POSITIVE LOGITS

 pursuit

0.33

 pursuits

0.29

 chase

0.27

 Purs

0.27

 shots

0.24

 pursued

0.24

 pursuing

0.23

 exchanges

0.23

 exchange

0.22

 shootout

0.22

Activations Density 0.054%

events related to violence or assaults involving weapons

No Comments

No Known Activations