INDEX

Explanations

dates in the format of year

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Configuration

jbloom/Gemma-2b-IT-Residual-Stream-SAEs/gemma_2b_it_blocks.12.hook_resid_post_16384

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

HuggingFaceFW/fineweb

Features

16,384

Data Type

float32

Hook Name

blocks.12.hook_resid_post

Hook Layer

Architecture

standard

Context Size

1,024

Dataset

Skylion007/openwebtext

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

wieś

-0.71

Jakie

-0.68

Przyp

-0.63

Stap

-0.61

Obr

-0.61

Welke

-0.59

Gdzie

-0.59

 gdyż

-0.59

Kako

-0.58

nasel

-0.58

POSITIVE LOGITS

 silikon

1.24

 kafe

1.18

 keramik

1.17

 panik

1.15

 seksi

1.15

 siff

1.09

 alkoh

1.08

 akut

1.07

 kaos

1.05

 katastro

1.05

Activations Density 0.110%

dates in the format of year

No Comments

No Known Activations