INDEX

Explanations

negativity and expressions of disappointment or dissatisfaction

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

STA

-0.07

HeaderCode

-0.07

aise

-0.07

ledon

-0.06

ampler

-0.06

 postav

-0.06

GuidId

-0.06

vro

-0.06

orno

-0.06

Ged

-0.06

POSITIVE LOGITS

0.08

Elect

0.07

mers

0.07

 from

0.07

son

0.07

is

0.06

 imper

0.06

inson

0.06

ious

0.06

Activations Density 0.181%

negativity and expressions of disappointment or dissatisfaction

No Comments

No Known Activations