INDEX

Explanations

references to being filled or crowded

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 0-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_0/width_16k/average_l0_105

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.0.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

raux

-0.55

thaler

-0.54

afa

-0.54

/***/

-0.53

vare

-0.52

["

-0.52

おきます

-0.51

ята

-0.51

pire

-0.50

 lancer

-0.49

POSITIVE LOGITS

 pack

1.03

 packed

0.99

 Packed

0.95

 crowd

0.93

 packing

0.89

 Crowd

0.89

 crowds

0.88

ArrowToggle

0.87

 Pack

0.87

Packed

0.85

Activations Density 0.175%

references to being filled or crowded

No Comments

No Known Activations