INDEX

Explanations

references to specific episodes in a series

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

start of turn user

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

 erfaren

-0.59

 erfaring

-0.56

 beroep

-0.55

 coscienza

-0.52

ambién

-0.52

 aprendizagem

-0.52

 voordeel

-0.52

 cartões

-0.50

 esperienze

-0.50

 econó

-0.50

POSITIVE LOGITS

 initComponents

0.60

setVerticalGroup

0.54

 reca

0.50

 recap

0.48

 jScrollPane

0.45

nice

0.45

 ripped

0.44

stdc

0.44

 חיצוניים

0.44

 nice

0.43

Activations Density 0.016%

references to specific episodes in a series

start of turn user

No Comments

No Known Activations