INDEX

Explanations

references to numerical quantities or time periods

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

times

-0.17

 Times

-0.17

-days

-0.17

æĢĿãģĦ

-0.16

anmar

-0.16

esis

-0.16

 Days

-0.16

Times

-0.16

 DAYS

-0.16

 dagen

-0.16

POSITIVE LOGITS

-plus

0.20

plus

0.16

 decades

0.16

 Plus

0.15

 editions

0.15

ownt

0.15

/releases

0.15

ÎµÏĦ

0.14

 fiscal

0.14

adu

0.14

Activations Density 0.037%

references to numerical quantities or time periods

No Comments

No Known Activations