INDEX

Explanations

references to legal terms and jurisdictions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 30-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L30R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.30.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

 Ø§ÙĦØ®Ø§ÙħØ³Ø©

-0.16

å¤ĩ

-0.15

ursed

-0.15

angel

-0.14

oodles

-0.14

zv

-0.14

 naken

-0.14

poser

-0.14

ovah

-0.14

Canceled

-0.14

POSITIVE LOGITS

 forum

0.29

Lug

0.27

lex

0.25

forum

0.25

 Forum

0.24

Forum

0.22

/forum

0.22

habit

0.20

 forums

0.20

 conflicts

0.20

Activations Density 0.008%

references to legal terms and jurisdictions

No Comments

No Known Activations