INDEX

Explanations

titles and headings in structured documents

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 15-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L15R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.15.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

zel

-0.19

âĤ¬↵

-0.15

ussy

-0.14

\Builder

-0.14

à¹ĩà¸Ķ

-0.14

okud

-0.13

MUT

-0.13

ä¸ĭåİ»

-0.13

POSITIVE LOGITS

 Â¿

0.18

zer

0.16

why

0.14

 seals

0.14

Why

0.14

 does

0.14

 strictly

0.14

endors

0.14

 Ø¢ÛĮØ§

0.14

 Does

0.13

Activations Density 0.232%

titles and headings in structured documents

No Comments

No Known Activations