INDEX

Explanations

words associated with military or governmental structure and processes

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

-0.18

acco

-0.15

cup

-0.15

 cups

-0.15

crawl

-0.14

 bÃ¬

-0.14

 Craw

-0.14

æĪĴ

-0.14

ekl

-0.14

ÏĦÏİÎ½

-0.14

POSITIVE LOGITS

 commission

1.34

 Commission

1.26

commission

1.16

Commission

1.14

 commissions

1.03

 commissioners

0.89

 commissioner

0.81

 Commissioners

0.79

 commissioned

0.78

 Commissioner

0.74

Activations Density 0.150%

words associated with military or governmental structure and processes

No Comments

No Known Activations