INDEX

Explanations

specific phrases or terms related to categorization, particularly involving attributes and classifications

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

or options, alternatives

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

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

Not in Any Lists

No Comments

Negative Logits

хьтан

-0.47

+#+

-0.44

MemoryWarning

-0.43

detectChanges

-0.43

 turnips

-0.43

habad

-0.43

 poch

-0.42

 Affiliations

-0.42

 onOptions

-0.42

 Torch

-0.41

POSITIVE LOGITS

それとも

0.63

or

0.59

 للاسماء

0.54

 Simplemente

0.47

 또는

0.45

 atau

0.43

 eller

0.43

 или

0.43

 irgende

0.42

 sebagainya

0.42

Activations Density 0.047%

specific phrases or terms related to categorization, particularly involving attributes and classifications

or options, alternatives

No Comments

No Known Activations