INDEX

Explanations

quantitative time-related terms and phrases

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 26-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L26R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.26.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

adar

-0.14

ë£

-0.14

ectl

-0.14

xor

-0.14

McL

-0.14

Ã©tico

-0.13

campo

-0.13

velt

-0.13

lech

-0.13

pora

-0.13

POSITIVE LOGITS

 after

0.39

 upon

0.37

-after

0.30

after

0.29

upon

0.29

 apÃ³s

0.28

 AFTER

0.28

 dopo

0.28

 Upon

0.27

 efter

0.27

Activations Density 0.037%

quantitative time-related terms and phrases

No Comments

No Known Activations