INDEX

Explanations

the neuron appears to activate on many words that do not share an obvious theme. More study is needed

oai_token-act-pair · gemini-2.0-flash

almost

np_max-act-logits · gemini-2.0-flash

New Auto-Interp

Configuration

google/gemma-scope-2b-pt-transcoders/layer_23/width_16k/average_l0_25

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.23.ln2.hook_normalized

Architecture

jumprelu_transcoder

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

BorderFactory

-0.47

tská

-0.46

koľvek

-0.45

extAlignment

-0.42

 Traducción

-0.41

koliv

-0.41

 acost

-0.41

/*---

-0.40

Clik

-0.40

おきます

-0.39

POSITIVE LOGITS

 nearly

3.09

nearly

2.91

 almost

2.89

 Nearly

2.75

almost

2.70

 Almost

2.69

Almost

2.64

Nearly

2.59

 quase

2.25

 hampir

2.16

Activations Density 2.564%

the neuron appears to activate on many words that do not share an obvious theme. More study is needed

almost

No Comments

No Known Activations