INDEX

Explanations

code-related words such as `fixes`, `breaks`, and `logic`.

oai_token-act-pair · gemini-2.0-flash

break

np_max-act-logits · gemini-2.0-flash

New Auto-Interp

Configuration

google/gemma-scope-2b-pt-transcoders/layer_24/width_16k/average_l0_37

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.24.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

évaluateur

-0.59

__);

-0.59

ArrowToggle

-0.58

Démographie

-0.56

/*

-0.56

Décès

-0.53

 Insee

-0.49

 Mihai

-0.49

Khan

-0.49

 Folly

-0.49

POSITIVE LOGITS

 break

2.56

break

2.44

 breaks

2.41

 Break

2.36

 broke

2.36

 broken

2.33

 breaking

2.28

Break

2.23

 Breaks

2.08

 BREAK

2.05

Activations Density 6.475%

code-related words such as `fixes`, `breaks`, and `logic`.

break

No Comments

No Known Activations