INDEX

Explanations

instances of the word "how" indicating inquiries or instructions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 23-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L23R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.23.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

à¸¸à¸ķ

-0.14

anni

-0.14

Äħd

-0.14

eniable

-0.14

ipo

-0.14

ÑħÐ¸

-0.14

 ÑĢÐ°ÑģÑħÐ¾Ð´

-0.14

uten

-0.14

../../../../

-0.13

orgt

-0.13

POSITIVE LOGITS

-to

0.19

itzer

0.19

 best

0.18

tog

0.17

esch

0.16

 properly

0.16

 proper

0.16

iza

0.15

to

0.14

Activations Density 0.035%

instances of the word "how" indicating inquiries or instructions

No Comments

No Known Activations