INDEX

Explanations

programming constructs and logic used in coding, particularly related to variable definitions and conditional structures

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

code, data, and queries

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

LookAnd

-0.85

 surla

-0.64

 betweenstory

-0.59

 pleaſure

-0.57

 kaarangay

-0.57

 ddelweddau

-0.54

 defaultstate

-0.54

Хьажоргаш

-0.52

oa̍t

-0.52

CppCodeGen

-0.51

POSITIVE LOGITS

awaiter

0.43

 information

0.38

 kans

0.38

info

0.38

sf

0.37

 Harris

0.37

 query

0.36

Query

0.36

 Kunt

0.36

uschen

0.36

Activations Density 0.028%

programming constructs and logic used in coding, particularly related to variable definitions and conditional structures

code, data, and queries

No Comments

No Known Activations