INDEX

Explanations

references to hyperbolic expressions and exaggerations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 29-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L29R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.29.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

ãĤĴè¦ĭãĤĭ

-0.15

âĢ¦"↵↵

-0.15

 Sesso

-0.14

eparator

-0.14

UDENT

-0.14

 ÅŁeyi

-0.13

oplan

-0.13

à¸¶à¹Ī

-0.13

 StringComparison

-0.13

ffect

-0.13

POSITIVE LOGITS

 anyone

0.43

 anybody

0.37

FT

0.34

0.33

 Anyone

0.31

 alert

0.29

 indeed

0.28

Anyone

0.27

 included

0.25

ft

0.24

Activations Density 0.541%

references to hyperbolic expressions and exaggerations

No Comments

No Known Activations