INDEX

Explanations

mentions of friends and family

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 24-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L24R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.24.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

cerebras/SlimPajama-627B

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Friend

-0.32

Friend

-0.30

friend

-0.29

 friend

-0.28

 friendship

-0.27

 Friendship

-0.24

 friendships

-0.23

 Friends

-0.21

_friend

-0.20

 Ð´ÑĢÑĥÐ¶

-0.20

POSITIVE LOGITS

 foes

0.31

 enemies

0.29

 acquaint

0.28

 neighbors

0.27

 Enemies

0.27

 neighbours

0.25

 colleagues

0.25

foe

0.24

 associates

0.24

 relatives

0.24

Activations Density 0.025%

mentions of friends and family

No Comments

No Known Activations