INDEX

Explanations

phrases indicating purpose or intention

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 21-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L21R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.21.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

vers

-0.15

laus

-0.15

uer

-0.14

etter

-0.13

isman

-0.13

ono

-0.13

ãģĩ

-0.13

 already

-0.13

rien

-0.13

ien

-0.13

POSITIVE LOGITS

 existence

0.19

Exist

0.17

 Exist

0.17

existence

0.16

Exists

0.15

ÏĦÎµÎ¹

0.15

ãĥ³ãĥĹ

0.15

Why

0.14

 ÙĪØ¬ÙĪØ¯

0.14

 Exists

0.14

Activations Density 0.060%

phrases indicating purpose or intention

No Comments

No Known Activations