INDEX

Explanations

Roman numerals

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

 Chwiliwch

-0.62

 noten

-0.61

 مرئيه

-0.60

StoreMessageInfo

-0.58

LEncoder

-0.56

 Paglinawan

-0.56

GARET

-0.54

UMBUS

-0.54

 distanciation

-0.54

glement

-0.54

POSITIVE LOGITS

 disagre

1.42

 inappro

1.40

 apprehen

1.40

 emphat

1.34

 uninten

1.31

 impra

1.30

 inconce

1.29

 unspeak

1.28

 increa

1.27

 depic

1.25

Activations Density 0.091%

Roman numerals

No Comments

No Known Activations