INDEX

Explanations

numbers and various forms of counts or measurements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 24-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L24R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.24.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

 Tues

-0.16

ieri

-0.15

engo

-0.14

piry

-0.14

wen

-0.14

 Riley

-0.14

Ere

-0.14

zeÅĦ

-0.14

Tuesday

-0.14

åįģ

-0.14

POSITIVE LOGITS

 Thursday

0.47

Thursday

0.47

Thu

0.39

 Thurs

0.39

0.38

0.36

0.35

Thu

0.34

0.33

Activations Density 0.154%

numbers and various forms of counts or measurements

No Comments

No Known Activations