INDEX

Explanations

language indicating that the author is working through a problem

oai_token-act-pair · gemini-2.0-flash

New Auto-Interp

Configuration

fnlp/Llama-Scope-R1-Distill/400M-Slimpajama-400M-OpenR1-Math-220k/L21R

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Hzfinfdu/SlimPajama-3B and open-r1/OpenR1-Math-220k

Features

32,768

Data Type

float32

Hook Name

blocks.21.hook_resid_post

Architecture

jumprelu

Context Size

1,024

Dataset

cerebras/SlimPajama-627B

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 accordingly

-0.08

so

-0.07

 Accordingly

-0.07

 thus

-0.07

 Thus

-0.06

 ìŀĪëĬĶëį°

-0.06

icha

-0.06

not

-0.06

 which

-0.06

",__

-0.06

POSITIVE LOGITS

 better

0.12

better

0.10

Rather

0.10

 rather

0.10

nor

0.10

 mejor

0.09

Better

0.09

 Better

0.09

 Rather

0.09

Actually

0.09

Activations Density 0.195%

language indicating that the author is working through a problem

No Comments

No Known Activations