INDEX

Explanations

sections and subsections in a formatted document

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 9-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L9R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.9.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

ÐµÑĤÑĥ

-0.15

usted

-0.15

gain

-0.15

ore

-0.14

Ð¾ÑģÐ»Ð°Ð²

-0.14

arb

-0.14

 Hancock

-0.14

 flexGrow

-0.14

ìļĶ

-0.13

 StartCoroutine

-0.13

POSITIVE LOGITS

æ·

0.19

 scop

0.14

ake

0.14

 Nationals

0.14

aken

0.14

ermann

0.13

Å¾en

0.13

ulpt

0.13

anz

0.13

Activations Density 0.003%

sections and subsections in a formatted document

No Comments

No Known Activations