INDEX

Explanations

names and identifiers, often related to people or groups

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

enor

-0.15

Ø¬ÙĪ

-0.14

eref

-0.14

.crm

-0.14

ilim

-0.14

raison

-0.13

 Holland

-0.13

illion

-0.13

jo

-0.13

redo

-0.13

POSITIVE LOGITS

raki

0.15

mall

0.15

__,__

0.14

ÏĢÎµ

0.14

anzeigen

0.14

hton

0.14

atables

0.14

ACING

0.13

 æ·±

0.13

Sa

0.13

Activations Density 0.005%

names and identifiers, often related to people or groups

No Comments

No Known Activations