INDEX

Explanations

patterns related to timestamps or date-related information

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 14-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L14R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.14.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

 Reason

-0.15

hell

-0.15

 Everett

-0.14

usband

-0.14

jiang

-0.14

 èĳ

-0.14

dpi

-0.14

uez

-0.14

ahoma

-0.13

aggio

-0.13

POSITIVE LOGITS

_drive

0.15

ffective

0.15

à¸Ļà¸²

0.15

à¸´à¸ģà¸²

0.15

vrier

0.15

ê¶Į

0.15

elli

0.14

#ab

0.14

ekim

0.14

ora

0.14

Activations Density 0.010%

patterns related to timestamps or date-related information

No Comments

No Known Activations