INDEX

Explanations

instances of the word "engage" and its variations, emphasizing active participation and interaction

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

Ø®Ø§ÙĨÙĩ

-0.20

 Ð¶Ðµ

-0.18

owie

-0.16

ahoma

-0.16

cedures

-0.16

ãĥĥãĤ¯ãĤ¹

-0.15

ÎĳÎļ

-0.15

Ð¾Ð²Ð°ÑĤÐµÐ»ÑĮ

-0.15

rello

-0.14

uraa

-0.14

POSITIVE LOGITS

Åĳ

0.16

/disable

0.16

uality

0.15

force

0.14

ä¿Ĺ

0.14

ÑģÐ¾Ðº

0.14

 mortar

0.13

/pass

0.13

prise

0.13

Activations Density 0.032%

instances of the word "engage" and its variations, emphasizing active participation and interaction

No Comments

No Known Activations