INDEX

Explanations

contractions and possessive forms indicating relationships or ownership

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 13-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L13R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.13.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

/or

-0.24

tti

-0.18

 latter

-0.17

ombat

-0.17

æĥħåĨµ

-0.17

äººåĵ¡

-0.16

’t

-0.16

culate

-0.15

ym

-0.15

’s

-0.15

POSITIVE LOGITS

nbsp

0.21

odore

0.19

 been

0.18

amp

0.18

ÂĿ

0.18

eresa

0.16

got

0.16

been

0.15

ÑįÑĤÐ¾Ð¼Ñĥ

0.15

apiro

0.15

Activations Density 0.345%

contractions and possessive forms indicating relationships or ownership

No Comments

No Known Activations