INDEX

Explanations

substrings and related operations

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

substring and peninsula

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

__["

-0.52

 Ordem

-0.46

}{*}{}

-0.46

 ligiloj

-0.45

 jeruk

-0.44

<strong>

-0.44

棣

-0.43

 vägen

-0.43

FIRE

-0.43

<blockquote>

-0.43

POSITIVE LOGITS

substring

1.48

Substring

1.13

 substring

1.09

 Paglinawan

0.66

substr

0.64

 substr

0.56

CGRectMake

0.54

InitVars

0.54

 addSubview

0.54

StoryboardSegue

0.54

Activations Density 0.007%

substrings and related operations

substring and peninsula

No Comments

No Known Activations