INDEX

Explanations

negations or phrases that indicate something is not true or not present

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

New Auto-Interp

Configuration

ctigges/pythia-70m-deduped__res-sm_processed/2-res-sm

Prompts (Dashboard)

32,768 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

32,768

Data Type

torch.float32

Hook Name

blocks.2.hook_resid_post

Hook Layer

Architecture

standard

Context Size

128

Dataset

EleutherAI/the_pile_deduplicated

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ķ

-2.77

ļ

-2.74

±

-2.74

ĻĤ

-2.71

ĨĴ

-2.71

»

-2.71

Ĵ

-2.71

Ļª

-2.68

®

-2.67

ĸ´

-2.66

POSITIVE LOGITS

oriously

2.56

 anymore

2.43

yet

2.24

 necessarily

1.90

necessarily

1.90

 enough

1.86

yet

1.81

allowed

1.81

tingham

1.73

etheless

1.73

Activations Density 0.328%

negations or phrases that indicate something is not true or not present

No Comments

No Known Activations