INDEX

Explanations

expressions of high praise or quality

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

 greatness

-0.16

Ø³Øª

-0.14

abi

-0.14

plevel

-0.14

theless

-0.14

elta

-0.14

elight

-0.13

unner

-0.13

emens

-0.13

POSITIVE LOGITS

0.30

-grand

0.29

sword

0.22

 dane

0.19

lest

0.17

(est

0.17

 deal

0.17

coat

0.17

ÏĤ

0.17

fully

0.17

Activations Density 0.048%

expressions of high praise or quality

No Comments

No Known Activations