INDEX

Explanations

references to locations and geographic features

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 28-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L28R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.28.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

 corridors

-0.18

æ°Ĺ

-0.15

 parts

-0.15

 Zones

-0.15

 zones

-0.14

 mach

-0.14

 Harbour

-0.14

WAYS

-0.14

eways

-0.14

fone

-0.13

POSITIVE LOGITS

 entrance

0.29

 edge

0.26

 mouth

0.25

 western

0.25

end

0.23

 eastern

0.22

western

0.22

 bend

0.21

entr

0.21

 southern

0.20

Activations Density 0.194%

references to locations and geographic features

No Comments

No Known Activations