INDEX

Explanations

mathematical and programming terms related to sequences and data structures

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

indent, lines, two, gamma, tunneling

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

 CreateTagHelper

-0.71

 مواليد

-0.69

msgTypes

-0.65

.*")]

-0.64

 HttpNotFound

-0.61

帖最后由

-0.61

RenderAtEndOf

-0.61

 незавершена

-0.60

IActionResult

-0.60

 createContext

-0.60

POSITIVE LOGITS

kw

0.39

---

0.30

bid

0.30

0.29

pat

0.29

 heard

0.29

 fault

0.29

 Escobar

0.29

sw

0.29

</b>

0.29

Activations Density 0.233%

mathematical and programming terms related to sequences and data structures

indent, lines, two, gamma, tunneling

No Comments

No Known Activations