INDEX

Explanations

uses of the word 'is' in various contexts

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ting

-0.07

ÑĨÐ°

-0.07

 prob

-0.07

 once

-0.06

 ones

-0.06

 following

-0.06

 figures

-0.06

carrier

-0.06

 pencil

-0.06

POSITIVE LOGITS

nameof

0.07

ViewById

0.07

'gc

0.07

ListOf

0.07

heck

0.07

_done

0.06

ãĢĪ

0.06

ä¸Ī

0.06

 ì²

0.06

istrovstvÃŃ

0.06

Activations Density 0.045%

uses of the word 'is' in various contexts

No Comments

No Known Activations