INDEX

Explanations

words and expressions related to uncertainty and negation, particularly regarding personal or political statements

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Configuration

Juliushanhanhan/llama-3-8b-it-res/blocks.25.hook_resid_post

Features

65,536

Data Type

float32

Hook Name

blocks.25.hook_resid_post

Hook Layer

Architecture

gated

Context Size

1,024

Dataset

Juliushanhanhan/openwebtext-1b-llama3-tokenized-cxt-1024

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

spor

-0.15

agraph

-0.15

Ð¶Ð½Ð¾

-0.15

unn

-0.14

mma

-0.14

anela

-0.14

urement

-0.14

wyn

-0.14

ÃŃny

-0.14

ycastle

-0.13

POSITIVE LOGITS

 Bark

0.18

0.17

çĦ¡ãģĹ

0.17

 sure

0.17

 against

0.17

oy

0.16

rade

0.15

 ready

0.15

 tá»ĭch

0.15

 someone

0.15

Activations Density 0.047%

words and expressions related to uncertainty and negation, particularly regarding personal or political statements

No Comments

No Known Activations