INDEX

Explanations

themes related to loss, harm, and risk

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 12-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L12R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.12.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

 ä¿Ŀ

-0.17

zza

-0.16

reate

-0.15

undy

-0.14

arna

-0.14

ä¿Ŀ

-0.14

oku

-0.14

glas

-0.14

 ÏĢÏģÎ¿ÏĤ

-0.14

 neighbour

-0.13

POSITIVE LOGITS

aky

0.15

 manip

0.14

alim

0.14

é¦ĸ

0.14

ìĭł

0.14

 Manip

0.13

guid

0.13

Logic

0.13

hardt

0.13

borough

0.13

Activations Density 0.003%

themes related to loss, harm, and risk

No Comments

No Known Activations