INDEX

Explanations

sections of text with high-frequency or repeated symbols

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 9-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L9R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.9.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

íħĲ

-0.16

edium

-0.16

iked

-0.16

malink

-0.15

glomer

-0.15

egas

-0.15

usercontent

-0.14

voje

-0.14

lus

-0.14

orer

-0.14

POSITIVE LOGITS

unc

0.15

 Batter

0.15

atter

0.14

zes

0.14

 Emer

0.14

 Matter

0.14

duct

0.14

DUCT

0.14

Wol

0.14

Alt

0.14

Activations Density 0.002%

sections of text with high-frequency or repeated symbols

No Comments

No Known Activations