INDEX

Explanations

words and phrases related to actions and performances

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

someone

-0.22

 someone

-0.21

 somebody

-0.20

ä¸Ģä¸ªäºº

-0.20

 sebuah

-0.20

 Someone

-0.19

 alguien

-0.18

si

-0.17

Someone

-0.17

htar

-0.15

POSITIVE LOGITS

 quite

0.23

 such

0.21

 Quite

0.20

quite

0.18

 SUCH

0.18

 amore

0.17

 somewhat

0.15

pend

0.14

arg

0.14

Activations Density 0.300%

words and phrases related to actions and performances

No Comments

No Known Activations