INDEX

Explanations

references to specific groups or individuals in context

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 24-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L24R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.24.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

iales

-0.16

Ĥ¹

-0.15

ouser

-0.15

ubes

-0.14

ibbon

-0.14

Ã¤s

-0.14

iasi

-0.14

<decltype

-0.13

antu

-0.13

Ø¯Ø§Ø¯

-0.13

POSITIVE LOGITS

 already

0.21

 need

0.18

 maybe

0.18

 otherwise

0.18

 require

0.17

 perhaps

0.16

require

0.16

 Already

0.16

 either

0.16

 Require

0.16

Activations Density 0.110%

references to specific groups or individuals in context

No Comments

No Known Activations