INDEX

Explanations

programming-related constructs and elements, particularly in code snippets

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

code statements and declarations

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

rungsseite

-0.69

RegressionTest

-0.64

 ErrIntOverflow

-0.63

 fieldNum

-0.62

ActionCreators

-0.61

ंदीखरीदारी

-0.60

 JpaRepository

-0.54

⟬

-0.54

VERTISEMENT

-0.54

 Infórmanos

-0.54

POSITIVE LOGITS

 surla

0.43

￡

0.35

 viſ

0.34

vis

0.31

आ

0.29

jsxFileName

0.28

pis

0.28

cas

0.28

 detec

0.27

 ſta

0.27

Activations Density 0.046%

programming-related constructs and elements, particularly in code snippets

code statements and declarations

No Comments

No Known Activations