INDEX

Explanations

hierarchical structures and subdivisions in categorization

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 18-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L18R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.18.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

 laid

-0.15

agua

-0.15

 è¾

-0.15

æĹıèĩªæ²»

-0.15

beer

-0.14

bih

-0.14

GU

-0.14

ãĥ¼ãĥ

-0.14

apult

-0.14

Gim

-0.14

POSITIVE LOGITS

/Sub

0.29

(Sub

0.28

/sub

0.24

.Sub

0.21

(sub

0.20

sub

0.20

=sub

0.20

ordinates

0.18

_Sub

0.18

ordinate

0.17

Activations Density 0.048%

hierarchical structures and subdivisions in categorization

No Comments

No Known Activations