INDEX

Explanations

references to structured sessions and training formats in a therapeutic context

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

one-on-one sessions or interviews

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

 obdob

-0.36

 médicas

-0.36

 sapin

-0.35

咪

-0.34

TestingModule

-0.34

 olsa

-0.33

 tabak

-0.33

 reemplazar

-0.33

 navideña

-0.33

 Bildung

-0.33

POSITIVE LOGITS

 tartalomajánló

0.61

 sessions

0.52

WebServlet

0.51

Guided

0.50

 Coaching

0.49

 coaching

0.48

Coaching

0.46

 defaultstate

0.46

 interview

0.46

+#+

0.46

Activations Density 0.012%

references to structured sessions and training formats in a therapeutic context

one-on-one sessions or interviews

No Comments

No Known Activations