INDEX

Explanations

phrases that structure relationships or chronological sequences

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 24-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L24R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.24.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

iken

-0.16

ysa

-0.14

 Easter

-0.14

sel

-0.14

iesen

-0.14

anan

-0.14

ast

-0.14

lein

-0.14

orig

-0.14

POSITIVE LOGITS

 history

0.18

 several

0.17

-urlencoded

0.17

 HISTORY

0.16

oyer

0.15

RuntimeObject

0.15

history

0.15

toa

0.14

NewProp

0.14

 History

0.14

Activations Density 0.029%

phrases that structure relationships or chronological sequences

No Comments

No Known Activations