INDEX

Explanations

references to various numerical or quantified concepts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 6-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L6R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.6.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

ots

-0.16

ows

-0.14

alchemy

-0.14

ÙģØ±

-0.13

ither

-0.13

rots

-0.13

ovel

-0.13

ovsky

-0.13

aman

-0.13

ade

-0.13

POSITIVE LOGITS

ulton

0.16

afone

0.15

Ãºn

0.15

/by

0.14

romo

0.14

-transparent

0.14

 parc

0.14

idor

0.14

 Sundays

0.13

_Abstract

0.13

Activations Density 0.042%

references to various numerical or quantified concepts

No Comments

No Known Activations