INDEX

Explanations

identifiers and key-value pairs related to user and organization attributes

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 14-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L14R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.14.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

ãĥ¼ãĥª

-0.18

uz

-0.16

edis

-0.15

heck

-0.15

OTH

-0.14

uan

-0.14

ÑģÑĭ

-0.14

 Vict

-0.14

ulus

-0.13

 ince

-0.13

POSITIVE LOGITS

olland

0.18

entiful

0.16

anness

0.15

arnation

0.15

#ac

0.14

marsh

0.14

chyb

0.14

ternet

0.14

ledon

0.14

ÑĢÐ°Ð²Ð°

0.14

Activations Density 0.146%

identifiers and key-value pairs related to user and organization attributes

No Comments

No Known Activations

identifiers and key-value pairs related to user and organization attributes

No Comments

No Known Activations