INDEX

Explanations

contractions, specifically "not"

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 6-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L6R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.6.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

cerebras/SlimPajama-627B

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

atern

-0.18

tain

-0.15

icina

-0.15

ropri

-0.15

attery

-0.14

raphics

-0.14

asn

-0.14

\Mapping

-0.14

Ð±Ð¸

-0.14

aidu

-0.14

POSITIVE LOGITS

oro

0.15

flo

0.15

0.14

 Lonely

0.14

 Proud

0.14

itia

0.14

ìĶ¨

0.13

Flo

0.13

izia

0.13

owi

0.13

Activations Density 0.078%

contractions, specifically "not"

No Known Activations