INDEX

Explanations

terms related to roles or identities, particularly in educational and professional contexts

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

 itself

-0.21

 ÐºÐ¾ÑĤÐ¾ÑĢÐ¾Ðµ

-0.16

azÄĥ

-0.15

ifecycle

-0.15

Â´t

-0.15

 koje

-0.14

lam

-0.14

ezi

-0.13

erras

-0.13

Ã¥n

-0.13

POSITIVE LOGITS

who

0.50

 whom

0.40

who

0.39

 whose

0.32

 quien

0.31

(s

0.29

Who

0.28

whose

0.27

Who

0.26

/client

0.25

Activations Density 0.370%

terms related to roles or identities, particularly in educational and professional contexts

No Comments

No Known Activations