INDEX

Explanations

references to sociopolitical concepts or terms related to history

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 21-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L21R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.21.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

utter

-0.17

utters

-0.15

https

-0.14

 builtin

-0.14

 Shooting

-0.14

á»©ng

-0.14

ÑģÑĸ

-0.14

idth

-0.13

 forces

-0.13

POSITIVE LOGITS

\Id

0.17

LOBAL

0.17

elow

0.15

AGED

0.14

ÙĦØ§ÙĨ

0.14

peÄį

0.14

//{{

0.14

 Albert

0.14

oser

0.14

Ã³Å¼

0.14

Activations Density 0.005%

references to sociopolitical concepts or terms related to history

No Comments

No Known Activations