INDEX

Explanations

titles and names related to literary or philosophical works and their authors

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 30-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L30R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.30.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

 whom

-0.15

atos

-0.14

bee

-0.14

ugal

-0.14

oyer

-0.14

 quienes

-0.14

 kteÅĻÃŃ

-0.14

ocre

-0.14

 clutch

-0.14

Ã©mon

-0.14

POSITIVE LOGITS

mem

0.19

 direct

0.18

 conduct

0.18

 Direct

0.17

 procure

0.17

PD

0.17

ami

0.17

PD

0.17

iffe

0.17

 foot

0.17

Activations Density 0.019%

titles and names related to literary or philosophical works and their authors

No Comments

No Known Activations