INDEX

Explanations

phrases related to denials, opposition, and intentions

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

 bonbons

-0.58

 createSlice

-0.57

 jette

-0.57

<%@

-0.57

 læng

-0.56

 tiens

-0.55

 appContext

-0.55

 rempliss

-0.54

 réuss

-0.54

 serai

-0.53

POSITIVE LOGITS

 whatsoever

0.56

 anything

0.53

nor

0.51

 wrongdoing

0.51

anything

0.49

any

0.49

Autoritní

0.48

 intend

0.47

dafx

0.46

 makro

0.46

Activations Density 0.338%

phrases related to denials, opposition, and intentions

No Comments

No Known Activations