INDEX

Explanations

conjunctions that indicate connections or relationships within a discussion

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 15-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_15/width_16k/average_l0_78

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.15.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Majefty

-0.87

Efq

-0.84

ualaikum

-0.82

 iconFacebook

-0.81

ſelf

-0.78

 pleaſure

-0.67

 houſe

-0.65

ſelves

-0.65

Witaj

-0.62

 THEM

-0.60

POSITIVE LOGITS

 although

0.98

 there

0.84

it

0.82

although

0.78

 aunque

0.73

0.70

Although

0.68

 Although

0.66

we

0.66

 this

0.65

Activations Density 0.365%

conjunctions that indicate connections or relationships within a discussion

No Comments

No Known Activations