INDEX

Explanations

phrases related to counting, ordering, and classification

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 30-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L30R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.30.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

entrant

-0.15

anzi

-0.15

.scalablytyped

-0.14

 siti

-0.14

strup

-0.14

 suce

-0.14

lectric

-0.14

ptic

-0.13

POSITIVE LOGITS

 greens

0.40

 whites

0.35

 Whites

0.35

 blacks

0.35

 doubles

0.33

 globals

0.33

 singles

0.33

 Greens

0.33

 blues

0.32

atives

0.31

Activations Density 0.226%

phrases related to counting, ordering, and classification

No Comments

No Known Activations