INDEX

Explanations

phrases indicating unresolved situations or actions yet to be completed

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 27-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L27R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.27.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

 already

-0.24

 Already

-0.23

already

-0.21

Already

-0.21

_already

-0.18

 dÃ©jÃł

-0.17

 earlier

-0.16

 continu

-0.16

 bereits

-0.15

 continuation

-0.14

POSITIVE LOGITS

 fully

0.32

 completely

0.23

 Fully

0.22

 entirely

0.22

 officially

0.22

å®Įåħ¨

0.21

Fully

0.21

official

0.21

fully

0.21

 official

0.20

Activations Density 0.048%

phrases indicating unresolved situations or actions yet to be completed

No Comments

No Known Activations