INDEX

Explanations

references to authority figures or influential individuals

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 28-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L28R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.28.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

cco

-0.16

ouz

-0.15

oard

-0.15

atrice

-0.15

ombok

-0.15

arie

-0.14

ÏĤ

-0.14

fix

-0.14

allah

-0.14

 Brilliant

-0.14

POSITIVE LOGITS

 sturdy

0.20

 functionalities

0.18

men

0.18

paced

0.16

 invalid

0.16

 Scotch

0.15

 turbulent

0.15

 honest

0.15

 gentle

0.15

ye

0.15

Activations Density 0.395%

references to authority figures or influential individuals

No Comments

No Known Activations