INDEX

Explanations

phrases that denote cause and effect or purpose

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 12-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L12R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.12.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

malink

-0.17

 withStyles

-0.16

ibold

-0.15

umba

-0.15

ician

-0.14

ithub

-0.14

lak

-0.14

oward

-0.14

oust

-0.14

sonian

-0.14

POSITIVE LOGITS

 chance

0.17

 chances

0.17

Chance

0.16

 Chance

0.15

ento

0.14

 deal

0.14

omi

0.13

aven

0.13

piration

0.13

 caus

0.13

Activations Density 0.127%

phrases that denote cause and effect or purpose

No Comments

No Known Activations