INDEX

Explanations

conditional phrases or suggestions aimed at the reader

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

ylon

-0.20

yh

-0.19

indre

-0.17

èĤ¡ä»½

-0.14

ollider

-0.14

ernel

-0.14

entiful

-0.14

Aware

-0.14

à¸ķà¸¥à¸Ńà¸Ķ

-0.13

uve

-0.13

POSITIVE LOGITS

 prefer

0.24

 fancy

0.23

 preference

0.20

 staying

0.18

 amore

0.18

 Prefer

0.18

 favor

0.18

 year

0.17

prefer

0.17

 love

0.17

Activations Density 0.077%

conditional phrases or suggestions aimed at the reader

No Comments

No Known Activations