INDEX

Explanations

references to existence and belief, particularly regarding concepts of evil, divinity, and societal structures

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 29-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L29R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.29.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

aepernick

-0.15

¼åĲĪ

-0.14

ÃŃn

-0.14

pok

-0.14

omu

-0.14

Ð¬

-0.14

ĵåĲį

-0.13

 localVar

-0.12

PÅĻÃŃ

-0.12

agma

-0.12

POSITIVE LOGITS

 exist

1.13

 exists

1.09

 existed

0.96

 existence

0.95

 Exist

0.93

exists

0.90

exist

0.87

 Exists

0.86

 existing

0.86

Exist

0.84

Activations Density 0.351%

references to existence and belief, particularly regarding concepts of evil, divinity, and societal structures

No Comments

No Known Activations