INDEX

Explanations

nouns and titles associated with positions of authority or significance

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 12-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L12R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.12.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

\common

-0.18

 Marsh

-0.15

å¸Ĥ

-0.14

duino

-0.14

illin

-0.14

Ã©ry

-0.13

Ø¨Ø§Ø´

-0.13

ingga

-0.13

olt

-0.13

 CITY

-0.13

POSITIVE LOGITS

0.17

 Veter

0.16

attern

0.15

ÄĽt

0.14

 Latter

0.14

ifice

0.14

Tam

0.14

ÙıÙĩ

0.13

Activations Density 0.183%

nouns and titles associated with positions of authority or significance

No Comments

No Known Activations