INDEX

Explanations

various forms of the word "Ninja."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 16-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L16R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.16.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

warts

-0.19

azar

-0.15

atos

-0.14

ird

-0.14

ÅĻez

-0.14

apest

-0.14

orr

-0.14

dip

-0.14

isher

-0.14

POSITIVE LOGITS

lake

0.16

ques

0.14

/callback

0.14

 Baxter

0.14

Sax

0.13

unate

0.13

Ã¡g

0.13

ges

0.13

 tear

0.13

LET

0.13

Activations Density 0.041%

various forms of the word "Ninja."

No Comments

No Known Activations