INDEX

Explanations

terms related to developers and development processes

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

developer documentation

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

newArrayList

-0.51

 Wain

-0.49

isTrue

-0.48

AddRange

-0.47

Nix

-0.47

 Informações

-0.45

tableFuture

-0.43

imb

-0.43

StrictEqual

-0.43

 Ababa

-0.42

POSITIVE LOGITS

 Developer

0.98

 developer

0.90

 Developers

0.87

developer

0.87

Developer

0.84

Developers

0.81

 developers

0.71

 programmer

0.71

 Programmer

0.67

 Coder

0.67

Activations Density 0.004%

terms related to developers and development processes

developer documentation

No Comments

No Known Activations