INDEX

Explanations

mentions of the word "Hus" or variations of it, which may refer to a person's name or specific context

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 26-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L26R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.26.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

.appendTo

-0.16

uzzi

-0.16

zes

-0.15

(strict

-0.15

tura

-0.15

commission

-0.14

zeitig

-0.14

ÑĤÑĢÐ¾

-0.14

ately

-0.14

ãģĹãĤĩ

-0.14

POSITIVE LOGITS

sein

0.27

sey

0.22

ain

0.22

Hus

0.19

bands

0.19

band

0.18

hus

0.18

Ã¡ng

0.18

aina

0.17

Band

0.17

Activations Density 0.006%

mentions of the word "Hus" or variations of it, which may refer to a person's name or specific context

No Comments

No Known Activations