INDEX

Explanations

punctuation marks, particularly commas and pipes, indicating lists and separations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 12-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L12R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.12.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

azine

-0.16

olin

-0.15

 AABB

-0.14

Kick

-0.14

ius

-0.14

-confirm

-0.13

@student

-0.13

 Ð¿ÑĢ

-0.13

omorphic

-0.13

inn

-0.13

POSITIVE LOGITS

.PropertyType

0.18

Ã¡ch

0.15

istro

0.15

taÅŁ

0.15

otta

0.15

AVOR

0.15

assi

0.15

.rmi

0.14

/rs

0.14

æ¡Ĥ

0.14

Activations Density 0.063%

punctuation marks, particularly commas and pipes, indicating lists and separations

No Comments

No Known Activations