INDEX

Explanations

references to traditional or conventional concepts, practices, or structures

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 14-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L14R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.14.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

ynn

-0.14

 à¤ī

-0.14

iso

-0.14

à¸µà¸¢à¸Ķ

-0.13

unas

-0.13

isp

-0.13

 militar

-0.13

LLU

-0.13

 particular

-0.13

ð

-0.13

POSITIVE LOGITS

oti

0.17

.esp

0.16

ily

0.16

zie

0.16

mente

0.16

/simple

0.15

wealth

0.15

Ã¶g

0.15

 Wisdom

0.15

-style

0.15

Activations Density 0.074%

references to traditional or conventional concepts, practices, or structures

No Comments

No Known Activations