INDEX

Explanations

repetitive expressions relating to the concept of "more."

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

æĽ´å¤ļ

-0.17

 some

-0.15

esa

-0.15

ÑģÑı

-0.14

ãģķãĤīãģ«

-0.14

 Ð½ÐµÐºÐ¾ÑĤÐ¾ÑĢÑĭÐµ

-0.14

asca

-0.14

 anymore

-0.13

 several

-0.13

astes

-0.13

POSITIVE LOGITS

 than

0.40

of

0.34

/all

0.28

-than

0.27

than

0.27

 THAN

0.25

 Than

0.22

Than

0.21

 importantly

0.19

 neÅ¾

0.19

Activations Density 0.128%

repetitive expressions relating to the concept of "more."

No Comments

No Known Activations