INDEX

Explanations

negative or dismissive phrases questioning someone’s actions or beliefs

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 26-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L26R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.26.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

 STILL

-0.18

 currently

-0.18

currently

-0.17

 Still

-0.17

still

-0.17

ãģĹãģ¦ãģĦãģ¾ãģĻ

-0.16

 still

-0.16

eum

-0.15

Still

-0.15

 presently

-0.15

POSITIVE LOGITS

 originally

0.29

 previously

0.26

 recently

0.23

 formerly

0.22

 Originally

0.21

 yesterday

0.20

 initially

0.19

actic

0.19

 invent

0.19

Didn

0.19

Activations Density 0.131%

negative or dismissive phrases questioning someone’s actions or beliefs

No Comments

No Known Activations