INDEX

Explanations

dialogue that involves advice and reflection on personal growth or accountability

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 30-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L30R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.30.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

fuck

-0.19

 fucked

-0.18

 fuck

-0.17

 FUCK

-0.17

 fucks

-0.15

 Fucking

-0.15

 Fuck

-0.15

 fucking

-0.15

 rapes

-0.14

Fuck

-0.14

POSITIVE LOGITS

 buddy

0.23

 partner

0.20

 buddies

0.20

 fellow

0.18

 boss

0.17

 brother

0.17

 amigo

0.17

 accomp

0.17

 intimidating

0.17

 mentor

0.17

Activations Density 0.158%

dialogue that involves advice and reflection on personal growth or accountability

No Comments

No Known Activations