INDEX

Explanations

punctuation marks and their frequencies

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

giving advice

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 20-gemmascope-res-65k

Configuration

google/gemma-scope-2b-pt-res/layer_20/width_65k/average_l0_114

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

65,536

Data Type

float32

Hook Name

blocks.20.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

Dziękuję

-0.86

IndentedString

-0.78

 насељу

-0.78

帖最后由

-0.76

]<<"

-0.74

"]);

-0.73

SharedDtor

-0.72

rungsseite

-0.72

ectoria

-0.72

Efq

-0.70

POSITIVE LOGITS

Tips

0.86

How

0.85

How

0.84

 Tips

0.83

 tips

0.80

If

0.79

If

0.74

Sometimes

0.70

Tip

0.69

 Whether

0.69

Activations Density 0.130%

punctuation marks and their frequencies

giving advice

No Known Activations