INDEX

Explanations

references to legal or religious concepts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 26-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L26R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.26.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

 âĹĦ

-0.18

 Starr

-0.18

aco

-0.16

unnel

-0.16

Ð°ÑĢÐ´

-0.15

à¤ªà¤¨

-0.15

UnitOfWork

-0.15

ÃŃÅ¾

-0.14

ardon

-0.14

EIF

-0.14

POSITIVE LOGITS

-S

0.17

imary

0.17

Sy

0.17

shaw

0.16

 conserv

0.16

Sy

0.15

jan

0.15

-s

0.15

 Jahres

0.15

-d

0.15

Activations Density 0.039%

references to legal or religious concepts

No Comments

No Known Activations

references to legal or religious concepts

No Comments

No Known Activations