INDEX

Explanations

unique identifiers and metadata, particularly in structured data like articles or reports

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 18-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L18R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.18.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

egie

-0.15

ingleton

-0.14

lehem

-0.14

Marsh

-0.14

emailer

-0.14

à¸§à¸Ķ

-0.14

Ú¯ÛĮ

-0.14

repid

-0.13

isma

-0.13

ÏģÎ¹

-0.13

POSITIVE LOGITS

ken

0.15

ane

0.15

enn

0.14

hod

0.14

otal

0.14

 gilt

0.14

woff

0.13

oro

0.13

 Survival

0.13

iya

0.13

Activations Density 0.090%

unique identifiers and metadata, particularly in structured data like articles or reports

No Comments

No Known Activations