INDEX

Explanations

references to significant buildings and landmarks

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 21-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L21R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.21.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

Ã¤ÃŁ

-0.16

ximity

-0.16

gren

-0.16

pora

-0.15

 microscope

-0.15

 hete

-0.15

ä»£çĲĨ

-0.15

 microscopic

-0.14

 parallel

-0.14

POSITIVE LOGITS

 tower

0.40

 towers

0.36

 Tower

0.35

 tallest

0.35

Tower

0.33

tower

0.32

 Towers

0.29

å¡Ķ

0.27

 taller

0.25

 tall

0.25

Activations Density 0.070%

references to significant buildings and landmarks

No Comments

No Known Activations