INDEX

Explanations

references to completion and success in tasks

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

programming code snippets

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

 Wikimedijinoj

-0.47

ioutil

-0.46

WritableDatabase

-0.41

ngdoc

-0.38

 bubble

-0.38

enment

-0.36

ensioni

-0.35

peteer

-0.34

postIndex

-0.33

 pleaſure

-0.33

POSITIVE LOGITS

transQ

0.55

 propOrder

0.50

 snippetHide

0.48

uxxxx

0.47

 ролях

0.42

fycat

0.42

 Signalez

0.42

 surla

0.42

يميديا

0.40

Activations Density 0.009%

references to completion and success in tasks

programming code snippets

No Comments

No Known Activations