INDEX

Explanations

numerous instances of positive and impactful attributes or achievements

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

olib

-0.16

-0.14

ropol

-0.14

th

-0.14

zel

-0.14

ifr

-0.13

 somewhat

-0.13

on

-0.13

 unchanged

-0.13

Ø¶Ø¹

-0.13

POSITIVE LOGITS

Î½Î¬

0.16

oted

0.15

awah

0.15

 DateTimeKind

0.14

ugi

0.14

errupt

0.14

imers

0.14

aÄį

0.14

isky

0.14

 INLINE

0.14

Activations Density 0.086%

numerous instances of positive and impactful attributes or achievements

No Comments

No Known Activations