INDEX

Explanations

references to geographical directions, particularly focusing on "north" and "south."

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

aldi

-0.15

ÏĦÎ±

-0.15

ibraltar

-0.15

odore

-0.15

bsites

-0.15

ity

-0.15

 Epoch

-0.15

aldo

-0.15

edl

-0.14

ãĤ£

-0.14

POSITIVE LOGITS

western

0.43

-east

0.38

-west

0.36

ward

0.34

bound

0.29

ern

0.28

wards

0.27

erner

0.25

west

0.25

-East

0.24

Activations Density 0.034%

references to geographical directions, particularly focusing on "north" and "south."

No Comments

No Known Activations