INDEX

Explanations

phrases relating to the nature and consequences of objects, particularly dangerous ones like guns, and societal issues surrounding them

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 29-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L29R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.29.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

jvu

-0.16

ÄŁer

-0.15

orners

-0.14

oku

-0.14

ementia

-0.14

ÑĪÐ¸Ð²

-0.14

afort

-0.14

Ì§

-0.14

estr

-0.14

mazon

-0.14

POSITIVE LOGITS

 alarm

0.29

 disappoint

0.28

 shock

0.27

 refresh

0.25

 delight

0.25

alarm

0.24

 distress

0.24

 disarm

0.24

 impress

0.24

 anything

0.23

Activations Density 0.473%

phrases relating to the nature and consequences of objects, particularly dangerous ones like guns, and societal issues surrounding them

No Comments

No Known Activations