INDEX

Explanations

phrases related to types or categories of various subjects

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 31-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L31R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.31.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

and

-0.20

 whilst

-0.19

 somehow

-0.18

 portions

-0.17

 with

-0.17

 whenever

-0.16

 hereby

-0.16

 among

-0.16

 amongst

-0.16

 some

-0.16

POSITIVE LOGITS

 thing

0.34

thing

0.26

 thinking

0.21

 Thing

0.20

 person

0.20

 activity

0.20

 arrangement

0.19

 behavior

0.19

ä¸ľè¥¿

0.19

 relationship

0.18

Activations Density 0.096%

phrases related to types or categories of various subjects

No Comments

No Known Activations