INDEX

Explanations

references to information inside brackets or encoded data

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

 Eighth

-0.19

 eighth

-0.18

enville

-0.16

imed

-0.16

fal

-0.15

redient

-0.15

Ø§ÙĨÙĩ

-0.15

ibri

-0.14

 gain

-0.14

 }};↵

-0.14

POSITIVE LOGITS

0.63

0.57

 nineteen

0.43

 nineteenth

0.39

 eighteen

0.38

0.37

0.33

Û±Û¹

0.32

 ninete

0.32

Û±Û¸

0.32

Activations Density 0.069%

references to information inside brackets or encoded data

No Comments

No Known Activations