INDEX

Explanations

specific numerical values and museum-related terms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 18-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L18R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.18.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´m

-0.17

lund

-0.16

æĹıèĩªæ²»

-0.15

ounc

-0.15

çĽ¤

-0.14

iali

-0.14

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

-0.14

esium

-0.14

·

-0.14

edly

-0.14

POSITIVE LOGITS

Ler

0.15

deb

0.14

lashes

0.14

 saddle

0.14

 Zimmer

0.13

 Proble

0.13

 Beginning

0.13

âĦĸâĦĸ

0.13

#.

0.13

èħ

0.13

Activations Density 0.245%

specific numerical values and museum-related terms

No Comments

No Known Activations