INDEX

Explanations

repeated references to the word "that."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 26-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L26R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.26.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

ãģĤãĤĬ

-0.23

ãģĤãĤĭ

-0.21

Ùĩ

-0.20

ãģĤãģ£ãģŁ

-0.20

-0.19

amp

-0.18

us

-0.17

sans

-0.16

sdale

-0.15

esktop

-0.15

POSITIVE LOGITS

ched

0.37

 particular

0.33

/th

0.30

 same

0.29

zelf

0.29

ching

0.25

 PARTICULAR

0.24

-ÑĤÐ¾

0.23

 SAME

0.23

 Ð¶Ðµ

0.22

Activations Density 0.184%

repeated references to the word "that."

No Comments

No Known Activations