INDEX

Explanations

phrases that express necessity and actions related to verification or consideration

oai_token-act-pair · gpt-4o-mini

New Auto-Interp

Configuration

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

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

open-r1/OpenR1-Math-220k

Features

32,768

Data Type

float32

Hook Name

blocks.15.hook_resid_post

Architecture

jumprelu

Context Size

1,024

Dataset

cerebras/SlimPajama-627B

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

oe

-0.07

locker

-0.07

kle

-0.07

vÄĽd

-0.07

SED

-0.07

 ÅĻÃŃj

-0.07

noinspection

-0.06

chwitz

-0.06

oor

-0.06

 personally

-0.06

POSITIVE LOGITS

igh

0.07

 MainAxisAlignment

0.07

ÑĢÐ°Ð¹

0.07

odal

0.06

inu

0.06

Æ¡

0.06

 Pare

0.06

 mutual

0.06

[js

0.06

agner

0.06

Activations Density 0.443%

phrases that express necessity and actions related to verification or consideration

No Comments

No Known Activations