INDEX

Explanations

actions involving carrying, holding, or lifting objects

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 7-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L7R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.7.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

alion

-0.19

athom

-0.18

 fills

-0.15

zik

-0.15

uos

-0.15

utilus

-0.15

 filling

-0.15

enal

-0.14

 consum

-0.14

uco

-0.14

POSITIVE LOGITS

anz

0.17

Ã¥n

0.17

 ÑĢÑĥÐºÐ°Ð¼Ð¸

0.15

ker

0.15

å³°

0.14

 hold

0.14

ROC

0.14

rophy

0.13

UNCH

0.13

anga

0.13

Activations Density 0.119%

actions involving carrying, holding, or lifting objects

No Comments

No Known Activations