INDEX

Explanations

elements related to bibliographic information and publishing

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 21-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L21R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.21.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

wu

-0.17

oder

-0.15

yan

-0.15

ç½²

-0.14

assin

-0.14

 spin

-0.14

oner

-0.14

icket

-0.14

iw

-0.14

oders

-0.13

POSITIVE LOGITS

ennis

0.17

cloth

0.16

 ÑĢÐµÐ¹

0.15

cco

0.15

REL

0.15

STAR

0.15

ç¢¼

0.15

osc

0.15

Î¾Î¹

0.15

//{{

0.14

Activations Density 0.005%

elements related to bibliographic information and publishing

No Comments

No Known Activations