INDEX

Explanations

specific programming constructs and elements

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

Follows "Q:" or abbreviation

np_acts-logits-general · gemini-2.0-flash

`AL` `STAND` `complaint` `tl`

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_20/width_16k/average_l0_71

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

-1.16

-1.09

-1.06

-0.97

-0.91

-0.90

-0.89

<eos>

-0.89

’

-0.88

POSITIVE LOGITS

 Савезне

1.79

LookAnd

1.72

 houſe

1.71

 Roskov

1.70

 pleaſure

1.68

 Houſe

1.60

 ویکی‌پدی

1.58

IUrlHelper

1.57

 perſon

1.57

 myſelf

1.56

Activations Density 0.677%

specific programming constructs and elements

Follows "Q:" or abbreviation

`AL` `STAND` `complaint` `tl`

No Comments

No Known Activations