INDEX

Explanations

instances of the word "more."

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

Not in Any Lists

No Comments

Negative Logits

æĽ´å¤ļ

-0.18

 otherwise

-0.16

 more

-0.16

ilight

-0.15

æ¯į

-0.14

 mais

-0.14

sworth

-0.14

alar

-0.14

Ú¯Ø±

-0.14

Century

-0.14

POSITIVE LOGITS

 than

0.54

-than

0.47

than

0.41

Than

0.39

 Than

0.36

 THAN

0.35

_than

0.34

 niÅ¼

0.31

tti

0.29

 importantly

0.29

Activations Density 0.181%

instances of the word "more."

No Comments

No Known Activations