INDEX

Explanations

references to specific people or characters from various contexts

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 15-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L15R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.15.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

cerebras/SlimPajama-627B

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

teil

-0.27

sh

-0.24

son

-0.23

ton

-0.22

sel

-0.22

teen

-0.21

te

-0.21

ten

-0.20

den

-0.20

gle

-0.19

POSITIVE LOGITS

etwork

0.23

eya

0.23

omics

0.23

etics

0.22

eye

0.20

ey

0.20

vironment

0.20

egative

0.19

ese

0.19

esse

0.19

Activations Density 0.427%

references to specific people or characters from various contexts

No Known Activations