INDEX

Explanations

instances of time-related descriptors or references

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

Followed by specific nouns (mission, year, example)

np_acts-logits-general · gemini-2.0-flash

temporal markers and following nouns

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 22-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_22/width_16k/average_l0_72

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.22.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

Filmographie

-0.68

 dumne

-0.63

]),

-0.62

 eorum

-0.60

 müm

-0.59

 său

-0.59

 riconoscimento

-0.58

<bos>

-0.57

CppMethod

-0.57

 tué

-0.54

POSITIVE LOGITS

 stuff

1.12

 shitty

1.00

 stupid

1.00

 thing

0.99

 crappy

0.99

 damn

0.98

 shit

0.92

 goddamn

0.91

 fucking

0.91

 crap

0.88

Activations Density 0.389%

instances of time-related descriptors or references

Followed by specific nouns (mission, year, example)

temporal markers and following nouns

No Comments

No Known Activations