INDEX

Explanations

font names and styles

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 26-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L26R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.26.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

ternal

-0.16

arcy

-0.15

acre

-0.15

xa

-0.15

lac

-0.15

 Freeman

-0.14

 Mein

-0.14

 Dickinson

-0.14

 Wahl

-0.14

occer

-0.14

POSITIVE LOGITS

regular

0.30

 Regular

0.29

 regular

0.29

 Bold

0.28

-bold

0.28

 bold

0.28

Regular

0.28

-Bold

0.28

Bold

0.26

ibold

0.26

Activations Density 0.005%

font names and styles

No Comments

No Known Activations