INDEX

Explanations

references to titles and articles in academic or formal contexts

oai_token-act-pair · gpt-4o-mini

New Auto-Interp

Configuration

fnlp/Llama-Scope-R1-Distill/800M-Slimpajama-0-OpenR1-Math-220k/L15R

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Hzfinfdu/SlimPajama-3B

Features

32,768

Data Type

float32

Hook Name

blocks.15.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

cerebras/SlimPajama-627B

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

.basic

-0.15

çºª

-0.14

 éĸ

-0.14

 surre

-0.14

.codec

-0.13

_pickle

-0.13

/ns

-0.13

Ã¼st

-0.13

POSITIVE LOGITS

 entitled

0.18

achel

0.18

 titled

0.17

0.15

äº

0.14

Bonjour

0.14

0.13

aday

0.13

ogens

0.13

omba

0.13

Activations Density 0.063%

references to titles and articles in academic or formal contexts

No Comments

No Known Activations