INDEX

Explanations

emotions expressed through character actions and thoughts

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

 blinked

-0.16

agit

-0.14

loff

-0.14

 insists

-0.14

 Census

-0.13

FK

-0.13

 criticised

-0.13

Ess

-0.13

aid

-0.13

Cry

-0.13

POSITIVE LOGITS

 pictured

0.23

 invol

0.22

 tens

0.20

win

0.20

 mentally

0.19

forced

0.18

sag

0.18

 felt

0.18

 forced

0.18

 supposed

0.18

Activations Density 0.221%

emotions expressed through character actions and thoughts

No Comments

No Known Activations