INDEX

Explanations

terms related to human actions and experiences

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 29-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L29R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.29.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

acht

-0.17

eel

-0.17

odnÃŃ

-0.16

 Äĳiá»ĥn

-0.15

rnd

-0.15

Ä±mÄ±zÄ±n

-0.14

iales

-0.14

bject

-0.14

ãĤ¿ãĥ¼

-0.14

.fp

-0.14

POSITIVE LOGITS

 bers

0.21

 Bers

0.19

-in

0.19

 ahead

0.19

-ahead

0.17

 unto

0.17

-on

0.17

hay

0.17

 astr

0.16

ogle

0.16

Activations Density 0.152%

terms related to human actions and experiences

No Comments

No Known Activations