INDEX

Explanations

temporal indicators such as specific days, weeks, and months

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

setContent

-0.17

 later

-0.16

çħ§

-0.15

æ³£

-0.14

æľĢè¿ĳ

-0.14

 oggi

-0.14

boru

-0.14

.setContent

-0.14

 ÑģÐµÐ³Ð¾Ð´Ð½Ñı

-0.14

 lately

-0.13

POSITIVE LOGITS

 ÙħÛĮÙĦØ§Ø¯ÛĮ

0.21

 alone

0.20

 afternoon

0.20

ago

0.20

alone

0.20

ino

0.19

 morning

0.19

INO

0.17

after

0.17

-end

0.16

Activations Density 0.110%

temporal indicators such as specific days, weeks, and months

No Comments

No Known Activations