INDEX

Explanations

occurrences of ordinal numbers and their respective positions in sequences or rankings

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 25-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L25R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.25.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

Meh

-0.16

arakter

-0.16

thren

-0.15

ãģıãĤī

-0.15

atori

-0.15

anter

-0.15

fi

-0.14

sed

-0.14

jit

-0.14

iring

-0.14

POSITIVE LOGITS

-round

0.26

 round

0.22

-string

0.18

 game

0.17

ê´Ģ

0.16

 straight

0.16

-game

0.16

round

0.16

reen

0.15

-year

0.15

Activations Density 0.038%

occurrences of ordinal numbers and their respective positions in sequences or rankings

No Comments

No Known Activations