INDEX

Explanations

terms related to appointments, vouchers, and warranties

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 24-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L24R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.24.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

cerebras/SlimPajama-627B

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

reh

-0.18

ëĭ¤ê°Ģ

-0.15

Ð»Ð¸Ñĩ

-0.15

zelf

-0.15

ãĢħ

-0.15

lices

-0.14

HIR

-0.14

lease

-0.14

erken

-0.14

contest

-0.14

POSITIVE LOGITS

able

0.23

-winning

0.23

holder

0.20

/lic

0.19

ee

0.19

ables

0.19

/license

0.18

 holders

0.17

ing

0.17

-taking

0.17

Activations Density 0.109%

terms related to appointments, vouchers, and warranties

No Known Activations