INDEX

Explanations

phrases that indicate specific numeric or categorical attributes

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

oad

-0.16

 kterÃ©ho

-0.15

pio

-0.15

ormsg

-0.15

moon

-0.14

ubber

-0.14

azio

-0.14

ÏĥÏĦÎµ

-0.14

pad

-0.14

ear

-0.13

POSITIVE LOGITS

 themselves

0.18

oints

0.16

agrams

0.15

rikes

0.14

stants

0.14

ishi

0.14

 Weber

0.13

opers

0.13

 Webb

0.13

linger

0.13

Activations Density 0.090%

phrases that indicate specific numeric or categorical attributes

No Comments

No Known Activations