INDEX

Explanations

instances of the word "find" in various contexts

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

find detection points

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

⟬

-0.91

<unused28>

-0.83

<unused51>

-0.83

<unused23>

-0.83

<unused42>

-0.83

<unused14>

-0.83

<unused52>

-0.83

<unused43>

-0.82

<unused41>

-0.82

[@BOS@]

-0.82

POSITIVE LOGITS

find

0.77

Find

0.66

 Find

0.63

Finder

0.57

FIND

0.55

finder

0.54

 finder

0.52

 Findlay

0.52

 Finder

0.49

 Encuentra

0.47

Activations Density 0.011%

instances of the word "find" in various contexts

find detection points

No Comments

No Known Activations