INDEX

Explanations

percentages and letters commonly associated with grading or classifications

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 9-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L9R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.9.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

flix

-0.16

dog

-0.15

Ã¤r

-0.14

Wax

-0.14

ABA

-0.14

ÙģØ±

-0.14

BC

-0.14

BC

-0.14

ando

-0.13

ablish

-0.13

POSITIVE LOGITS

(æ°´

0.15

igated

0.15

 GLint

0.14

 Habit

0.14

 Haram

0.14

habit

0.14

 capitalize

0.14

iesz

0.14

LPARAM

0.13

Activations Density 0.003%

percentages and letters commonly associated with grading or classifications

No Comments

No Known Activations