INDEX

Explanations

symbols or characters in different encoding formats

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 31-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L31R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.31.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

erli

-0.18

 getItemCount

-0.17

ÐµÐ¼Ð°

-0.15

 Belt

-0.14

ARGER

-0.14

ÏģÎ±

-0.14

/Foundation

-0.14

riÃ¨re

-0.14

 Grinder

-0.14

 ----------------------------------------------------------------------------↵

-0.13

POSITIVE LOGITS

 save

0.31

save

0.27

 saving

0.26

-save

0.25

 saved

0.25

 Save

0.25

 SAVE

0.24

 saves

0.24

.save

0.22

 Saves

0.22

Activations Density 0.007%

symbols or characters in different encoding formats

No Comments

No Known Activations