INDEX

Explanations

phrases and references related to goals, decisions, and progress in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 14-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L14R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.14.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

 continued

-0.15

emento

-0.15

hai

-0.14

jang

-0.14

 continu

-0.14

 continuing

-0.14

Ã¡li

-0.14

 Ø±Ø¯

-0.14

 Armour

-0.14

ather

-0.13

POSITIVE LOGITS

yet

0.35

yet

0.29

Yet

0.24

Yet

0.24

 fully

0.23

 henÃ¼z

0.22

 Fully

0.21

ãģ¾ãģł

0.20

Fully

0.17

 belum

0.17

Activations Density 0.143%

phrases and references related to goals, decisions, and progress in various contexts

No Comments

No Known Activations