INDEX

Explanations

references to preferred locations and activities associated with them

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

 italian

-0.16

 discrepan

-0.16

Italy

-0.15

 Italians

-0.15

Bea

-0.15

.uk

-0.15

ioni

-0.14

spb

-0.14

 Boeh

-0.14

POSITIVE LOGITS

 Latin

0.51

Latin

0.43

 Colombia

0.40

 Peru

0.40

 Colombian

0.38

 Ecuador

0.38

 Chile

0.37

 LATIN

0.37

Bog

0.36

 Bolivia

0.36

Activations Density 0.472%

references to preferred locations and activities associated with them

No Comments

No Known Activations